Lecture Notes in Logistics

Series Editors: Uwe Clausen Michael ten Hompel Robert de Souza

Michaelten Hompel
JakobRehof
OliverWolf Editors

Cloud
Computing
for Logistics
Lecture Notes in Logistics

Series editors
Uwe Clausen, Dortmund, Germany
Michael ten Hompel, Dortmund, Germany
Robert de Souza, Singapore, Singapore
More information about this series at http://www.springer.com/series/11220
Michael ten Hompel Jakob Rehof

Oliver Wolf
Editors

Cloud Computing
for Logistics

123
Editors
Michael ten Hompel Oliver Wolf
Fraunhofer-Institut fr Materialfluss und Fraunhofer-Institut fr Materialfluss und
Logistik IML Logistik IML
Dortmund Dortmund
Germany Germany

Jakob Rehof
Fraunhofer-Institut fr Software- und
Systemtechnik ISST
Dortmund
Germany

ISSN 2194-8917 ISSN 2194-8925 (electronic)

Lecture Notes in Logistics
ISBN 978-3-319-13403-1 ISBN 978-3-319-13404-8 (eBook)
DOI 10.1007/978-3-319-13404-8

Library of Congress Control Number: 2014956490

Springer Cham Heidelberg New York Dordrecht London

Springer International Publishing Switzerland 2015
This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or
dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt
from the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained
herein or for any errors or omissions that may have been made.

Printed on acid-free paper

Springer International Publishing AG Switzerland is part of Springer Science+Business Media

(www.springer.com)
Preface

Logistics is currently in a complexity trap: Globalization is constantly increasing,

the degree of networking is growing exponentially, batch size one is already a
reality. The volume of logistics data is increasing by a factor of 1000 per decade.
At the same time, the stability of logistics systems and processes is decreasing.
Therefore, there is no doubt about the necessity of flexibilization of the classic
Supply Chain Management. Cloud Computing is now helping to cut the Gordian
knot: The Cloud as a system environment has the potential to design complexity to
a certain extent controllably and processes more flexibly. It is about far more than
new applications and markets. It is about the fundamental redesign of logistics
systems. Cloud Computing has the ability to vitally change the kind of controlling
and organizing of these systems.
The Fraunhofer innovation cluster Cloud Computing for Logistics is an
extraordinary example to show the power that can be developed by logistics and
informatics as drivers of innovations. The Logistics Mall with several awards in the
meantime created a market place and an infrastructure that help providers and users
to test the new possibilities of Cloud Computing in a practice-oriented way. For its
users the Logistics Mall makes sure that Cloud Computing in the consciousness of
logistics is more than a cloud. Cloud Computing is understood as solution
pushing forward logistics companies.
Logistics is surely one of the most important and most innovative application
industries for information technologies (IT): it has internalized and pushed forward
the topic of Cloud Computing because it realized very early the chances connected to
it. However, the potential of Cloud Computing is by far not yet exploited as clearly
pointed out in the position paper of Bundesvereinigung Logistik (BVL) e.V.
Logistics and IT as innovation drivers for the business location GermanyThe new
leading role of logistics in information technology. The Fraunhofer Institutes for
Material Flow and Logistics (IML) and for Software and System Technology (ISST)
meet these challenges and have bundled their competencies for this purpose in the
Fraunhofer Innovation Centre for Logistics and IT (FILIT) since the middle of 2014.

v
vi Preface

Sponsored by:
Ministry of Innovation, Science and Research of the State of North Rhine-
Westphalia
Fraunhofer Society (Fraunhofer-Gesellschaft zur Frderung der angewandten
Forschung e.V.)

December 2014 Michael ten Hompel

Jakob Rehof
Oliver Wolf
Contents

Logistics Software Systems and Functions: An Overview

of ERP, WMS, TMS and SCM Systems . . . . . . . . . . . . . . . . . . . . . . . 1
Andreas Nettstrter, Tim Geien, Markus Witthaut,
Dietmar Ebel and Jens Schoneboom

Logistics MallA Cloud Platform for Logistics . . . . . . . . . . . . . . . . . 13

Damian Daniluk and Bernhard Holtkamp

Empirical Qualitative Analysis of the Current Cloud

Computing Market for Logistics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
Maren-Bianca Wolf and Jonas Rahn

The Logistics MallAn IT-Architecture for Logistics-as-a-Product . . . 45

Bernhard Holtkamp

Business Apps Meet the Challenge of Covering Continually

Changing Logistics Requirements. . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
Julian Eggemann, Jens Leveling and Norbert Wei

Seamless Interoperability in Logistics: Narrowing

the Business-IT Gap by Logistics Business Objects . . . . . . . . . . . . . . . 77
Martin Bhmer, Michael Schmidt and Norbert Weienberg

Challenges of Cloud Business Process Management. . . . . . . . . . . . . . . 119

Iryna Bochon, Volker Ivens and Ralf Nagel

vii
Logistics Software Systems and Functions:
An Overview of ERP, WMS, TMS
and SCM Systems

Andreas Nettstrter, Tim Geien, Markus Witthaut, Dietmar Ebel

and Jens Schoneboom

Abstract Software systems are a crucial part in current logistics processes. The
software market for Enterprise Resource Planning, Warehouse Management,
Transport Planning and Supply Chain Management systems and the systems
themselves are complex. The following article gives an overview about denitions
and typical functions of these systems as well as a short introduction to current
logistics and technological requirements.

Keywords Logistics software WMS ERP TMS SCM Enterprise resource

planning

Warehouse management

Transport planning Supply chain

management

1 Introduction

Logistics analyses and models the flow of goods, people and information through
nearly all economic systems and provides recommendations for their design and
implementation. The main focus in research is therefore on conguration,

A. Nettstrter (&)
T. Geien
M. Witthaut
D. Ebel
J. Schoneboom

Fraunhofer Institute for Material Flow and Logistics, Dortmund, Germany
e-mail: [email protected]
T. Geien
e-mail: [email protected]
M. Witthaut
e-mail: [email protected]
D. Ebel
e-mail: [email protected]
J. Schoneboom
e-mail: [email protected]

Springer International Publishing Switzerland 2015 1

M. ten Hompel et al. (eds.), Cloud Computing for Logistics,
Lecture Notes in Logistics, DOI 10.1007/978-3-319-13404-8_1
2 A. Nettstrter et al.

Fig. 1 Connection and information flow between ERP, WMS, TMS and SCM

organization, control and regulation of networks and flows with the objective to
optimize these with economic, environmental and social objectives in mind.
Logistics software systems are dened within this paper as systems (meaning
software and data) to support the above mentioned logistics tasks, including ade-
quate planning and forecasting tools along the entire logistics supply chain. This
includes all logistics areas from inhouse logistics over transport logistics to supply
chain management. Typical software systems in these areas are Enterprise Resource
Planning Systems (ERP), Warehouse Management Systems (WMS), Transportation
Management Systems (TMS) and Supply Chain Management systems (SCM).
The general connection and information flow between ERP, WMS, TMS and
SCM systems is shown in Fig. 1. SCM systems are not limited to a single company,
but used within cross company supply networks. The ERP systems are usually
company based just like WMS and TMS. The differences between SCM and ERP
systems become more and more blurred. Many of todays ERP systems support
multi-site capabilities and offer cross-site planning and processing functions and
can therefore be used for the management of supply chains.

2 Denition of Systems

2.1 ERP

In the 1990s ERP systems with a large range of functionality for several company
departments were mainly used in large companies. Small and midsize companies
worked with expert systems for resource planning, production (PPS), Warehouse
Management (WMS) and nancial accounting. Fully integrated ERP systems
became affordable for small and midsize businesses with decreasing cost for
Logistics Software Systems and Functions 3

Fig. 2 Key and extended functionality of typical ERP systems

hardware and software. In recent years ERP systems gained a more and more
functional expansion and many former separated expert systems were integrated
into the ERP systems. An ERP system today is a tool for comprehensive planning,
coordination and management of companywide tasks. The advantage here lies in
the efcient use of all existing resources of a company (e.g. capital, human
resources and information). Apart from logistic applications (such as inventory
management or disposition) ERP systems offer tools for almost all functions of a
company, such as nance, accounting, controlling, manufacturing, research and
development. Modern ERP systems must be able to adapt to dynamic business
processes and they must change into flexible software solutions using open tech-
nical standards. Figure 2 shows the basic and extended functionality of standard
ERP systems.
The market for ERP systems includes more than 200 vendors, but not everyone
offers a self-developed ERP system. The market for ERP systems can be divided in
the following categories:
Software manufacturers develop and sell their own products and are responsible
for all technological and functional properties.
Resellers sell ERP systems as partners of software manufacturers to customers.
Resellers are strong in market segments, where the software manufacturers do
not distribute and sell their own ERP system.
Suite vendors are also working with software manufacturers and enhance the
functional scope of the ERP system by industry-specic modules or functions.
4 A. Nettstrter et al.

The advantage for the user is a consistent environment with all functions
integrated.
Implementation partners support the customer during the implementation of the
ERP system with services.
There are also software manufacturers, resellers and suite vendors, who act as
implementation partners at the same time.

2.2 WMS

Since the 1970s software systems for warehouses and inventory management are
used to support logistic processes. Originally, these were systems to manage
quantities and (storage or bin) locations as well as the relations in between them [1].
Additional functionalities could manage means of transport. Hence, warehouse
management systems started as stock management systems.
After the major cost saving potentials of logistics in general and warehouses in
particular were discovered, warehouse management systems no longer were sys-
tems to only manage the stock of a warehouse, but became integrated software
systems with superior optimization and management functionalities.
Thus, warehouse management today describes the control, monitoring and
optimization of complex warehousing and distribution systems. Next to the ele-
mentary functionalities of managing quantities and storage locations or controlling
and scheduling the means of transport, comprehensive methods and instruments to
supervise the systems status belong to the scope of a modern warehouse man-
agement system as well as a selection of operating and and optimization strategies.
Consequently, the business of warehouse management systems is the operation and
optimization of on-site warehousing systems.
Because customers constantly ask for further functionalities to realize the possi-
bilities of cutting the costs for their warehouses, vendors of warehouse management
systems constantly enlarge the scope of their software systems. As a consequence,
warehouse management systems nowadays increasingly offer functionalities which
have their origin in enterprise resource planning software, supply chain management
software or transport management systems such as an entire order fullment, the
support of all processes in between receiving and shipping and comprehensive
information systems and control panels. Further functionalities which more and more
often belong to the standard of a modern warehouse management system are
the planning of tours and routes, the support of vendor managed inventories, or
the support of billing and value added services in case of multi-client scenarios.
These core and additional functionalities (see also Fig. 3) are completed by
optional subsystems (add-ons) for communication with the picker (e.g. Pick-
by-Light, Put-to-Light, Pick-by-Voice controls) or systems to identify the products
(e.g. RF readers, handhelds).
Logistics Software Systems and Functions 5

Fig. 3 Key and extended functionality of typical WMS

2.3 TMS

Before, IT-based transport management systems have been applied externally and
transport related logistics was usually carried out with the aid of forwarder-
customized software of the transport and logistics service providers. In case of more
complex problems route-planning software was additionally applied. On the part of
the shippers external logistics was covered by simple functional elements within
ERP systems in the past.
Due to the increasing complexity of global and closely linked logistic chains at
the end of the 1990s, the necessity of complex systems for the management of these
complex transport chains with all their functional requirements began to
develop. These systems are called transport management systems (TMS).
The resulting tasks cover the planning and optimization of procurement and
distribution structures under the consideration of e.g. costs or time restrictions, the
planning of multimodal transport chains or the optimization of the delivery trans-
ports as well as the control and monitoring of the realization of the resulting
6 A. Nettstrter et al.

Fig. 4 Key and extended functionality of typical TMS

transport processes. The integration of the mobile units (transport means, load
handling accessory equipment, etc.) by telematics is also an important element for
control and monitoring.
TMS in general allow the planning, control and monitoring as well as the
optimization of transport networks and logistic chains. Elementary functional elds
of these systems are order management, scheduling, transport planning and opti-
mization, tracking and tracing as well as fleet and resource management, see Fig. 4.
The market for transport management systems is already existing for several
years. The enterprises providing transport management solutions can be divided in
different groups:
Pure transport management software developers are providing their own prod-
ucts or as extensions in combination with other software elements. The reason
could be found in a specialization of the software solution for a specic industry on
the one hand. On the other hand it could also be found in the size of the developing
enterprise. Therefore, the developing enterprises are cooperating with providers
group of the sales partners selling the product to denite customers.
Logistics Software Systems and Functions 7

The sales partners are more service-orientated and solely focusing the sale of
transport management solutions. Furthermore, there the so-called resellers who are
also providing consultancy in addition to the sale of software. This allows the group
to provide a wide range of different productsfrom the pure solution for road up to
the combination of the transport meansrespectively with a different range of
functionsthe reason for being very flexible and the ability to offer a wide range.
The providers can be described as a combination of the types mentioned above.
Those are often medium-sized or big enterprises having an own development
division as well as a well-organized sales division. For TMS this group is the focus,
it is developing and implementing its software solutions directly and additionally
making customer consulting.
Also the provided product range and the functional spectrum are often more
versatile at big providers than at small enterprises. Furthermore, this group of
providers is able to develop further software solutions beyond the system interface
e.g. ERP and WMSthus also able to provide a complete package for bigger
customers. Until now the number of enterprises also providing TMS in addition to
ERP and WMS is limited. However, considering current extensions some of the
ERP providers are increasingly moving towards TMS. In most cases an increasing
range of TMS functions causes signicantly increasing costs of the total system.
Due to the complexity of the products offered at the market both small and big
enterprises will nd a suitable offer of transport management systems.
In transport-intensive industries the application of software solutions is of sig-
nicant importance. In the past years the demand for transport solutions and thus
for transport management systems has been considerably increasing. This shows
that the application of TMS is nearly unavoidable. In this respect the objectives of
the users are scarcely differing. High saving potentials of the resources as well as a
cost-optimal and efcient order management are more and more the goal. Differ-
ences in the requirements specication are mainly to be found in the size of the
enterprise and the related horizon of the order spectrum.
The key competences of the TMS providers cover the planning and optimization
of procurement and distribution structures with minimum costs which could be
both, economic and ecologic ones. The systems, however, have also differences in
view of the functional range and the technological base. Especially bigger providers
show the trend of offering more strategic planning elements in their software. There
are also transsectoral functionsas for example freight cost management or
planningprovided with only few systems.
The complexity of transport management systems is often based in the inte-
gration of manifold frame conditions and restrictionsbe that a legal regulation or
the versatile and different demands of the customers for these systems. This and the
worldwide opening of the transport networks are also considerably promoting the
competitive pressure on part of the TMS -providers so that those are especially
forced to continuously develop their systems.
The market development of the past years therefore shows the focus on con-
stantly new requirements that the transport managements systems have to fulll.
The recent past indicated a trend in the modeling of multimodal transports. Even for
8 A. Nettstrter et al.

complex European-wide or global transport chains with different transport means

schedulers and planners are expecting a support for the planning and control,
similarly to road-related transports being used to. In this context the ecologic
assessment of especially multimodal transports and logistic chains also plays an
important role.

2.4 SCM

Supply Chain Management (SCM) deals with inter-organisational production and

logistics networks. The notion supply chain is somehow misleading because SCM
addresses not only inbound logistics but the business activities associated with all
phases of satisfying a customers demand [2]. Furthermore, instead of single
supply and/or distribution chains SCM covers arbitrary networks. Hence, the notion
Value Net Management would be more suitable but is not used in the eld. Sim-
ilarly, the organisational coverage of SCM is differently dened in the eld and the
scientic literature. A point of view is that SCM integrates all logistical planning
and execution tasks for sourcing, manufacturing, and delivery of a single plants
products. Another extreme opinion especially common in academic papers
encompasses all sourcing, transport, manufacturing and distribution activities in
network of separate enterprises spanning from raw material supplier to the nal
customer and including return and after-sales services as well.
The different SCM conceptions lead to a variety of denitions for SCM software.
An extreme position is that SCM has to cover all logistical planning and execution
task in cross-enterprise production and logistics networks. However, there is no
such SCM software solution available on the SCM software market. Even though
there are software vendors offering solutions for the entire planning and execution
tasks, these systems cannot be applied across enterprise boundaries. For a pragmatic
approach we recommend a conception of SCM as the design, planning and exe-
cution of both intra- and inter-organisational material and information flows.
Consequently, we propose a task model encompassing all activities with IT support
for the management of production and logistics networks (see Fig. 5). The software
systems can be further structured into three main groups. The rst group serves for
the design of networksSupply Chain Design (SCD). SCD-Systems cover issues
like location planning including, transport system design and the dimensioning of
warehouse in respect to customers demand and production capacities. Supply
Chain Planning (SCP) includes planning tasks for a given network. Based on
customers demand, overall sourcing, manufacturing and distribution plans are
generated. Furthermore, some SCP-Systems offer features for the collaborative
demand and capacity planning with customers and suppliers. Available-to-Promise
functionality is used for due dating customer enquiries on a supply chain level. The
last SCM-functionality group is called Supply Chain Execution. With the exception
of supply chain event management, SCE-functions overlap with the functionality of
ERP, TMS and WMS software (see Fig. 5). An initial version of this SCM task
Logistics Software Systems and Functions 9

Fig. 5 Key and extended functionality of typical SCM systems

model was developed especially for a comparative market study on supply chain
management software in 1998/1999. A major revision was made in 2003. The
resulting model comprises the typical modules of SCM software and is further
based on the internationally established Supply Chain Operations Reference-model
(SCOR).

3 System Requirements

The current evolution of logistics system shows that modern information technol-
ogy will be more and more coupled to all elds of logistics. Only through this close
coupling the complexity of logistics can be handled and already today there are just
a few elds of logistics left, which are able to work efciently without the use of IT
systems. The software solutions used are as numerous as are their heterogeneous
areas of applications. However, there are still many barriers both, from the users
perspective as well as from the perspective of the software provider, contrary to the
use of IT in logistics. The next section will give a brief overview about user
requirements to logistics software systems. Additionally, many technological trends
will nd their way into logistics IT systems which are covered in Sect. 3.2.
10 A. Nettstrter et al.

3.1 Users Requirements

Based on a survey made last year with 208 logistics software users and providers
[3] ten requirements emerged as the most important. The requirements were divided
into two groups, system-specic logistics requirements and global technological
requirements. The logistics requirements are the following:
Increase of punctuality in the planning and control of logistic processes for
production and transportation of goods. This could be solved through the usage
of more precise scheduling methods and modern identication systems.
Creation of robust logistics systems which are able to continue operation also in
case of errors or unforeseen incidents. Possible solutions that are expected of the
software systems are, sufcient dimensioning of resources, integrated simulation
and calculation of alternative scenarios.
Increase of interoperability between different systems, e.g. for data or document
exchange. This requirement is closely coupled to the technological requirement
of open or standardized interfaces.

3.2 Technological Requirements

Technological requirements often mentioned in the survey for future logistics IT

systems are briefly described below:
Flexible: The system must be expandable for future changes and new (currently
unknown) requirements. The ability to respond to unforeseen events is known as
the changeability of systems.
Individual: Custom requirements and additions can easily be implemented.
On-Demand: The software should be available directly, without a long-term
implementation phase or the need for additional IT infrastructure.
Open/standardized interfaces: The software should use public and open inter-
faces for information exchange so that other systems can be connected without
major integration effort.
Low cost: The capital and operating costs must be minimized.
Intuitive HMI: The graphical human-machine-interface should be self-explan-
atory and understandable.
Usable on mobile devices: The software should be accessible also from outside
the company via mobile devices, like smartphones or tablets.
These requirements make it inevitable for the providers of logistics IT to keep an
eye on current technological trends which can offer possible solutions to meet these
requirements.
Logistics Software Systems and Functions 11

References

1. ten Hompel, M., Schmidt, T.: Warehouse Management. Organisation und Steuerung von Lager-
und Kommissioniersystemen. Springer, Berlin (2008)
2. Supply Chain Council: SCOR 10.0 Model Reference. Supply Chain Council, Cypress (2010)
3. ten Hompel, M. (ed.): IT in der Logistik. DVV Media Group, Hamburg (2012)
Logistics MallA Cloud Platform
for Logistics

Damian Daniluk and Bernhard Holtkamp

Abstract Common characteristics of all logistics processes are individuality and

dynamically changing requirements of the customers business. Cloud computing
enables new business models to provide highly individual IT services that t the
needs of logistics customers. After outlining logistics specic cloud service
requirements and the results of a study about the acceptance of cloud computing in
logistics domain this paper presents the Logistics Mall, an approach for a domain
specic cloud platform for the trading and usage of logistics IT services and
logistics processes.

Keywords Logistics mall Cloud computing Cloud platform Business process
as a service

1 Introduction

In 2011, the overall global logistics market represented a total volume of approx-
imately 981 Billion [1]. According to BVL (recognized german logistics insti-
tution), in Germany logistics is the third largest economy branch with an market
volume of approximately 235 Billion in 2014 and a work force of 2.9 Million.
The German logistics market, as typical for the logistics domain, is characterized
through small and medium size enterprises with little or no IT capacities and
competences besides operating their own IT resources.

D. Daniluk (&)
Fraunhofer Institute for Material Flow and Logistics IML,
Joseph-Von-Fraunhofer-Str. 2-4, 44227 Dortmund, Germany
e-mail: [email protected]
B. Holtkamp
Fraunhofer Institute for Software and Systems Engineering ISST,
Emil-Figge-Str. 91, 44227 Dortmund, Germany
e-mail: [email protected]

Springer International Publishing Switzerland 2015 13

M. ten Hompel et al. (eds.), Cloud Computing for Logistics,
Lecture Notes in Logistics, DOI 10.1007/978-3-319-13404-8_2
14 D. Daniluk and B. Holtkamp

Trade and industry consider logistics as a cost factor and as a factor of com-
petitiveness at the same time. As a consequence there is a growing trend for
outsourcing and contract logistics to benet from scale effects and from synergies.
This way the market for logistics services has developed from classical transport
transshipmentwarehousing services towards a growing market of more individual
services which are offered by specialized logistics service providers. There is also a
trend this services become more complex increasingly. One of the reasons for this is
the development in the e-commerce sector. In Germany, purchases over the internet
result in more than 100 million additional packages per year. Each of these pack-
ages is ordered individually and on demand, picked, packed, transported, distrib-
uted and delivered. At the same time, the number of items increases exponentially.
This means that customers of logistics service providers today require individual-
ized logistics services with:
a flexible and broad service spectrum,
individual logistics processes and value-added services,
transparency of costs and performance,
short-term contracts.
These requirements increase the following problems, evident for many logistics
service providers:
It usually lacks sufcient investment funds (or the willingness to invest) for the
expansion of IT.
Often it lacks IT expertise and availability of human capacity to operate the
necessary IT infrastructure adequately.
The development of new IT components and their integration into the existing
IT landscape is time-consuming.
Many logistics companies neither have adequate IT skills and capabilities, nor
the capital needed to close the gap between requirements and the status quo in the
logistics IT. In addition, there is growing pressure on the logistics sector through
legislative changes. Moreover, street charges and the need for electronic customs
clearance imply the use of new IT services, e. g. for control of driving and rest
times. The same applies to the introduction of tracking and tracing for continuous
condition monitoring of goods.
In this context cloud computing is a very interesting technology to face the
outlined challenges. The rest of the paper is organized as follows. In this chapter the
requirements of logistics on application domain specic cloud services are dis-
cussed. Section 2 summarizes the results of a study about the acceptance of cloud
computing within the logistics domain in Germany carried out by the Fraunhofer
Institute for Material Flow and Logistics IML. Section 3 provides an overview of
the Logistics Mall development. The paper closes with a summary of results and an
outlook to future work.
Logistics MallA Cloud Platform for Logistics 15

2 Cloud Computing in the Logistics Domain

Cloud computing is considered as a megatrend which will have an impact on the

usage of information technology in all application domains. Forrester forecasts that
the global market for cloud computing will grow from $40.7 billion in 2011 to more
than $241 billion in 2020. The total size of the public cloud market will grow from
$25.5 billion in 2011 to $159.3 billion in 2020 [2].
For a common understanding we adopt the working denition of NIST [3] that
denes cloud computing as a model for enabling convenient, on-demand network
access to a shared pool of congurable computing resources (e.g., networks,
servers, storage, applications, and services) that can be rapidly provisioned and
released with minimal management effort or service provider interaction. This cloud
model promotes availability and is composed of ve essential characteristics, three
service models, and four deployment models. The basic characteristics are on-
demand self-service for consumers, broad network access, resource pooling to serve
multiple consumers using a multi-tenant model, rapid elasticity of resources and
metering capabilities for service provision. The identied service delivery models
are software-as-a-service (SaaS) where consumers use a providers application
running on a cloud infrastructure, platform-as-a-service (PaaS) where a consumer
can deploy an application on a cloud infrastructure using a providers tools and
platform, and infrastructure-as-a-service (IaaS) where a consumer can run arbitrary
software on a provided cloud infrastructure.
Recently, big IT companies have invested many million dollars in the devel-
opment of cloud technology and in the provisioning of commercial cloud offers.
The span reaches from IaaS solutions like Amazons Elastic Compute Cloud [4] to
PaaS offers like Microsoft Windows Azure [5] and SaaS offers like Salesforce CRM
[6]. A common characteristic of these offers is their generic nature. They all con-
centrate on broad usage scenarios but do not provide business relevant features.
These cloud offers are called horizontal clouds. In their report on the future of
cloud computing [7] a European expert group identies the provisioning of
application area specic cloud services, also called vertical clouds, as a signi-
cant business opportunity. The German excellence cluster EfciencyCluster Lo-
gisticsRuhr (www.efzienzcluster.de) shares this view and has declared the
development of a domain specic logistics cloud a strategic target.

2.1 Requirements on Logistics Services Out of the Cloud

Logistic applications often have specic characteristics that are implied by the
requirements of the industry. By shifting logistics applications to the cloud, these
characteristics have to be considered in order to provide a proper replacement for
the existing customizable software solutions. The essential requirements that must
be considered in the context of logistics specic cloud services include:
16 D. Daniluk and B. Holtkamp

Integration of customer-specic peripherals

Realization of short response times of logistics software
Ensuring data security and transparency
Timely implementation of highly-customized business processes at low cost
Many of these requirements are not just technical in nature. However, they lead
to technical requirements on cloud computing on closer inspection.

2.1.1 Integration of Customer-Specic Peripherals

Typical software systems used in intralogistics include warehouse management

systems (WMS). In addition to the basic functions including the management of
quantities, bin locations and transport orders a warehouse management systems also
includes extensive means of controlling the system states and implements a variety
of operating and optimization strategies. Typical peripherals that have to be inter-
faced with a WMS include [8]:
Forklift and handheld terminals: These components are wireless computers for
mobile applications such as for forklifts or for manual operation. They are used
for timely and easy collection of data, for example as part of the picking of
products.
Barcode reader (infrared scanners) and transponder reader (RFID scanner):
Scanners are wired (e.g. USB) or connected wirelessly to a PC. The captured
data is processed by the WMS.
(Label) printers: Printers serve in the context of WMS as a medium for the
creation of goods receipt documents, picking lists or goods issue papers. Special
label printers can write adhesive labels which are attached as forwarding labels
on pallets.

2.2 Realization of Short Response Times

In logistic applications, compliance with upper limits of the response time of the
software is very important. For example, if it takes too much time to update a user
dialogue in applications like high performance picking, the employee cannot
continue with his work (e.g. scanning of products) and is hindered in his work
routine. This results in a reduction of productivity. To prevent such occurrences,
cloud applications should be high-performant, i.e., the response time of the appli-
cation for user input must be as low as possible so that user productivity is not
affected. Particularly, the use of cloud computing may not lead to an obvious
reduction in the performance of the user-frontend.
A particular challenge is the control of real time material flow systems. In the
eld of continuous conveyor systems guaranteed response times of less than 10 ms
Logistics MallA Cloud Platform for Logistics 17

are standard. This can only be realized through the use of a material flow control
systems on site, which may be coupled with applications in the cloud that are less
time-critical. On the other hand autonomous designed conveyor modules increas-
ingly allow a decoupling of real-time operations and thus the direct combination of
cloud and material flow control.

2.2.1 Information and Data Security

Information and material flow in the logistics domain have reached a very high
level of complexity. On the one hand this results from the variety of involved
entities (suppliers, customers, means of transport, machinery, goods, IT systems,
etc.). Another reason for this are multiple dependencies between the involved
entities. Logistics is responsible for the planning, control and optimization of
material and information flow that goes beyond company boundaries. As part of the
cooperation between companies more and more short-term co-operations are
formed, which depend on the core competencies of the companies. The tendency of
this development is increasing.
The logistics of today is not imaginable without modern IT, wireless data
transmission and mobile devices. Decentralization and particularly networking play
an increasingly important role in logistics. This development leads to new
requirements for existing security concepts. Especially in the logistics domain, the
customers of logistics service providers have a legitimate, strong interest to protect
their data (e.g. throughput data, sales data, and master data). Also, the use of mobile
devices (GPS, telematics, RFID readers, PDAs, smartphones and laptops) is
ubiquitous in logistics. Especially here it is important to develop concepts that
enable the safe use of such mobile hardware in the cloud.

2.2.2 Individual Business Processes, Implementation Timeframes

and Costs

In particular, the time factor blocks the development of flexible and innovative IT-
based logistics solutions in several ways. The implementation time of today
logistics systems is very long in proportion to the period of their utilization. Thus,
business models and strategies in industry and commerce currently have a typical
life cycle time of 23 years. The adequate modeling of (logistics) business pro-
cesses often claims a realization time of 610 months. If additionally there has to be
implemented a new, efcient IT solution for providing the required logistics ser-
vices, then the implementation period extends to an average of 12 up to 24 months.
The operating time of such strategic investments is an average of 35 years. After
this time period normally the IT solution has to be adjusted because of new or
changed requirements. This leads to an economically problematic relationship
between planning and realization expenses and usage times [9].
18 D. Daniluk and B. Holtkamp

This problem can be mitigated by IT components that can adapt to new

requirements simply and quickly by dynamic combination to support more complex
processes. For a cloud computing infrastructure this means that it must support the
loose coupling of functional services in the form of autonomous IT services. A
complete business process can be built up on this basis by choreography or
orchestration of IT services [9].

2.3 Acceptance of Cloud Computing in Logistics

The acceptance of cloud computing in logistics was analyzed in the current market
study Cloud Computing for Logistics of the Fraunhofer Institute for Material
Flow and Logistics IML in Dortmund [9]. In this analysis providers of logistics IT
services and potential users from the elds of logistics service providing, commerce
and industry were interviewed. The goal was to nd out the conditions under which
the logistics and IT managers are willing to consider and use cloud computing
approaches for critical logistics applications such as warehouse management sys-
tems (WMS).
The results of the study speak for themselves: the degree of acceptance of
logistics solutions in the cloud is very high. Already today the majority (56 %) of
corporate leaders can imagine to rent and run logistics software an external servers.
This shows that the market is open-minded for the usage of logistics software from
the internet and there exists a great potential for this form of software distribution.
The following section describes the Logistics Mall, a platform that allows the
purchase and usage of logistics software from the internet based on cloud
computing.

3 Logistics Mall

3.1 Vision

Common characteristics of all logistics business processes are individuality and

dynamically changing requirements of the customers business. A process has to be
adapted to every new or changed requirement. Consequently, the underlying IT
infrastructure is also subject to these frequently changing requirements.
Today a typical logistics process runs in a heterogeneous IT environment. Such
systems are mostly standard software like SAP, Oracle, or more specialized
logistics software systems like warehouse management systems (WMS) or pro-
duction planning systems (PPS). All these systems do not entirely meet the
requirements of logistics customers regarding short contract duration, pay-per-use
accounting and the provisioning of individual IT services. The majority of the
Logistics MallA Cloud Platform for Logistics 19

features of standard software is rarely used or not used at all, but has to be paid for,
due to a monolithic architecture. Customizing software to adapt to new or changed
requirements involves high efforts, costs and risks. Nevertheless the solutions
flexibility for upcoming adaptations is not increased.
The introduction of cloud computing is a new opportunity to deliver different IT
services over the internet. The idea of this IT paradigm is the abstraction of the
underlying software and hardware. Based on this, users do not have to manage and
administrate the physical hardware or software they are using. An additional
advantage is that software can be acquired on-demand and paid per use [10].
The concept of providing services over a cloud encompasses three different
stakeholders. Basically these are the operator of the cloud computing environment
(CCE) being responsible for its administration, providers offering their applications
or IT services and customers purchasing and using the services [11].
Regarding the dynamically changing requirements cloud computing provides the
opportunity for a logistics customer to rent IT services only as long as they are
needed. Furthermore, using a CCE to provide and use services enables customers
and logistic service providers to focus on their key business by outsourcing the IT
infrastructure to the cloud. Both customers and logistic service providers do not
have to establish and administrate an internal IT infrastructure and only require a
connection to the internet for interacting with the cloud. Additionally, pay-per-use
accounting is another advantage for the customers. For providers cloud computing
is a new opportunity to gain greater market relevance and design new offers by
connecting their products with services provided by other independent providers
(according to the slogan The whole thing is more than the sum of its components)
[12, 13].

3.2 Concept

The main idea of cloud computingAnything as a Servicecan be adapted to

source complete logistics business processes, designed by connecting single IT
services using an appropriate tool that is executed in the browser. Therefore, an
adequate process modeling methodology is necessary to offer logistics customers an
opportunity to model individual processes themselves. This service delivery model
is dened as Business Process as a Services (BPaaS). There are contributions that
use BPaaS as abbreviation to name models that have the general idea of outsourcing
of existing processes into the cloud [14]. Here, more specically, the unique feature
of this model is that both modeling and process execution are part of the cloud
platform. This idea is the basis of the Logistics Mall, a development of the
Fraunhofer Innovation Cluster Cloud Computing for Logistics [15], that focusses
on the modeling and execution of processes that are built out of several IT services
and that can be offered using a cloud platform. Most of the today available
approaches are too complex to be used by logistics customers that have minor
software engineering skills. On the one hand the vision of the Logistics Mall is to
20 D. Daniluk and B. Holtkamp

create a methodology that is capable to deliver business processes that can be

deployed automatically for execution. On the other hand the logistician may not be
overburdened by technical details.
The main idea of the Logistics Mall and the offering of BPaaS is visualized in
Figs. 1 and 2. Todays monolithic software solutions are replaced by small, dedi-
cated IT services of different service providers which can be combined to superior
services that support the individual business processes of the logistician.
In addition to the offering of BPaaS solutions the Logistics Mall is also able to
provide classic IT services corresponding to the Software as a Service (SaaS) cloud
model. The main reason for this is to address as many software providers as
possible to be able to offer IT systems that support a broad range of logistics
processes. For an IT service that is offered in conjunction with BPaaS the technical

Fig. 1 Todays monolithic software solutions and their interaction within a business process

Fig. 2 The Logistics Mall as a cloud platform offers IT-services that can be combined to support
individual logistics business processes
Logistics MallA Cloud Platform for Logistics 21

cloud

physical processes

Fig. 3 Concept and stakeholders of the Logistics Mall

requirements are much higher than for an IT service that is offered only in as SaaS
in the manner of ASP (application service providing). The reason for this is that
BPaaS requires the IT service to use the same business object data model so that in
the context of a modeled process IT services of different service providers have a
common understanding of the data that is communicated between them. A detailed
description of this data model and how it supports logistics business processes and
reduces the business-IT gap is provided by [16]. Moreover the IT services in
context of BPaaS should have a consistent look and feel of the GUI when they are
used to support one business process. This requirements force service providers to
undertake broad modications on their existing IT services to be offered on the
Logistics Mall. In contrast, providers have to do only slight modications to their
web-based IT services to offer them as SaaS products on the Logistics Mall plat-
form (cf. Sect. 3.4).
Part of the concept of the Logistics Mall (Fig. 3) is that IT developers can offer
and operate IT services in the cloud. Another target group is companies that have
the function of integrators. As logistics process designers they combine IT services,
which are compatible to be used in context of BPaaS, to more widespread IT
solutions that support complete logistics business processes. The third group is the
logistics customers, who order and consume the IT solutions offered by the
Logistics Mall.
Conceptually the Logistics Mall also supports the integration of physical ser-
vices. Such serviced could be offered on the Logistics Mall platform by logistics
service providers. Examples of these include transport services, which encompass a
transport of goods from a source to a destination.
22 D. Daniluk and B. Holtkamp

Overall, the Logistics Mall is a comprehensive approach of an integrated plat-

form that offers PaaS for IT developers and SaaS/BPaaS for logistics customers.
The usage model behind the Logistics Mall for both is its operation as a public
cloud or as a private cloud. Commercial operation of the Logistics Mall is available
to the public at http://www.logistics-mall.com.
The design of the Logistics Mall is based on two pillars. The rst pillar is the
shopping frontend in the form of the mall marketplace (MMP). The MMP provides
a publicly available online shop for the offered products. The second pillar of the
Logistics Mall is a platform that allows the operation and usage of the IT solutions
rented on the MMP by the logistics customer. After ordering a product in the MMP
it is provided on the so called Customized Access Framework (CAF).

3.3 Components of the Logistics Mall

The following section describes some essential elements of the Logistics Mall from
a functional viewpoint. An architecture overview about the Logistics Mall is pre-
sented in [17].

3.3.1 User Management

User management in the Logistics Mall is realized as a cross-component that is used

at both the MMP and the CAF. Every user account of the Logistics Mall is assigned
to an explicit company, in the wording of the Logistics Mall called client. A
distinction is made between three types of clients, namely the operator, the pro-
viders and the customers. The operator has an integral access to the Logistics Mall
through which he can access all the functions necessary for administration. Pro-
viders and customers get access to the Logistics Mall as part of a registration
process. During this process, every company gets assigned a unique client number.
Besides user name and password as user-specic identity information the unique
client number is required for the MMP login. For the CAF login the client number
is not necessary because the CAF is provided client-specic and is accessed via a
unique ULR assigned to the client.
The user management of the Logistics Mall allows a client to create arbitrary
other user accounts for users of his company and to assign permissions to each of
these accounts. The permissions relate to functions of both MMP and CAF. For
example, a user of type customer that has all possible permissions can access both
MMP and CAF and use all available functions and rented IT services with their full
functional range.
The assignment of permissions to users is done via roles. Each role aggregates a
set of permissions. On the one hand roles may have been created in the user
management of the Logistics Mall and then refer to functions that are provided in
MMP and CAF. On the other hand roles may have been exported by the IT services
Logistics MallA Cloud Platform for Logistics 23

the customer has rented and refer to permissions within the specic IT service. The
user information and the roles of a logged in client can be requested by the IT
services using a specic Logistics Mall interface. This concept called single sign on
is realized across the entire Logistics Mall. A user only needs to login once and then
can use all the features and IT services that he is allowed for by the permissions of
his user account.
Thus, the user management built into the basic infrastructure of the Logistics
Mall is a central element, which simplies the usability of the Logistics Mall and
reduces the organizational overhead.

3.3.2 Product Management

To capture new IT services in the Logistics Mall, there is a dialogue-based wizard,

which queries all necessary product information from the IT service provider in a
structured form. The registered provider can access the wizard on the MMP.
In the rst step of the wizard, the general product information such as product
name, version, classication and text, logos and flyers for product description are
queried. In the second step the specication of the accounting model is done for the
product. A distinction is made here between three accounting types: one-off costs,
flat-rate costs (in regular intervals recurring costs) and transaction-based costs
which support pay per use accounting. Transactions can be seen as clearly identi-
able events that written by an IT service using a given logging interface of the
Logistics Mall. Through the analysis of these logs a transaction-based billing is
possible. A combination of these three accounting types is called pricing model. A
pricing model may be assigned to the product itself or to any associated product
option separately.
The specication of product options and associated pricing models is the third
step of the dialog based wizard. Product options describe additional services that
can be ordered by the customer together with the product. An example for such
additional services is functional add-ons for the product. The product options can be
aggregated into so called option groups. In addition, it can be specied whether a
customer may choose only one option or several options from an option group as
part of the ordering process.
In the fourth step of the wizard information about contractual design and
accounting of the product is queried. These details include the minimum contract
period and cancellation period. In the nal fth step, a preview of the product is
presented. The provider can see how its product will be presented in the public
MMP product catalog. The Logistics Mall workflow requires a product to be cer-
tied and released by the operator before it is presented in the MMP product catalog
and can be rented by customers.
Generally, the product management of the Logistics Mall provides a compre-
hensive instrument with which the product and its associated accounting model can
be described individually by the provider of an IT service.
24 D. Daniluk and B. Holtkamp

3.3.3 Logistics Process Designer

The logistics process designer (LPD) is the central tool of the Logistics Mall that
allows the modeling of IT solutions that support the individual business processes
of a logistician. To be ready for business, the LPD supports a modeling method-
ology with the following characteristics:
The LPD can be used by a logistics expert without detailed IT knowledge.
A process consists of available IT services only. Modeling an arbitrary process
just to notice it cannot be run by the IT services available is not practical.
The LPD supports in its modeling process different functional granularity of IT
services and the CCE. Forcing IT service providers to only implement services
with specied, xed granularity limits the distinction of IT services and hampers
the growth of a broadly diversied range of IT services covering all ranges of
functionality demanded by customers.
The IT services within the modeled logistics process are exchangeable.
On premise IT systems that are not operated in the cloud can be integrated into
modeling by specic interfaces. By this, customers have the opportunity to
migrate to the cloud step by step, still using existing IT systems and connecting
them with IT services provided by the cloud.
A detailed description of the approach of the LPD is discussed in [18].

3.3.4 Reporting

Each client of the Logistics Mall gets access to a client-specic documents area. At
adjustable, regular intervals, reports are stored that contain diverse dynamic data:
Tickets: statistics about tickets, that have been sent using the helpdesk of the
Logistics Mall
Hardware utilization: details about utilized CPU time and memory utilization
originated from IT service operation in the CAF
Transactions: Data of the transactions the IT services in the CAF have written
for pay per use accounting
Reports are customized specically for the client types operator, provider and
customer. For example, the transactions report for end customers contains the
number of transactions logged for each rented product. For the provider the same
report contains transaction numbers for each of his products and for each of his
customers.
The reporting component of the Logistics Mall allows clients to trace data that
occurs in operation and is relevant for business and accounting.
Logistics MallA Cloud Platform for Logistics 25

3.3.5 Helpdesk

The Logistics Mall incorporates an integrated ticketing system, which enables the
processing of support requests. The 1st level support for end users is carried out by
the operator of the Logistics Mall. With this the operator represents the rst instance
to which customers direct their support requests. If the processing of a request is not
possible by the operator because the problem-solving requires a deeper under-
standing of an application, the operator forwards the request to the provider of the
application. The provider thus takes the 2nd level support and can reply to requests
of customers using the integrated ticketing system of the Logistics Mall.
Since the ticketing system belongs to the basic infrastructure of the Logistics
Mall, users can access it from both MMP and CAF. This provides a platform that
can be used to discuss and solve customers problems.

3.4 Interfaces for Application Integration

To make applications available in the Logistics Mall, interfaces are provided for the
application integration that must be implemented by the provider. The application
integration is largely based on Web service interfaces, as this will ensure that the
interfaces can be addressed through almost all modern programming languages.
Subsequently the main interfaces of the Logistics Mall are shortly described. These
are related to classic IT services corresponding to the Software as a Service (SaaS)
cloud model. Interfaces of IT services, which are used in the context of BPaaS in
the Logistics Mall, are described in [19].
IT service GUI integration: two different interfaces are provided for the inte-
gration of IT service GUIs. Supported are already web-enabled applications. A
simple integration is provided with the help of so-called inline frames. They
allow that a browser window containing the IT service GUI can be shown as
component of another browser window. In addition, application GUI can be
integrated via portlet technology.
Transaction-based accounting: to implement the transaction-based accounting
for each application the relevant business transactions are queried within the
product management wizard (cf. Sect. 3.3). During operation the application
then reports its transactions through an interface to the Logistics Mall. The
transactions are logged for accounting purposes.
Connection of external peripherals: for reliable connection of the typical
peripherals in logistics environment a special interface is provided by the
Logistics Mall. This interface allows an application to establish a secure con-
nection to the periphery.
File storage: applications often write log les, reports or exchange data using a
shared le. These les also must be stored in a cloud environment like the
Logistics Mall. For this reason, Logistics Mall provides an interface through
26 D. Daniluk and B. Holtkamp

which IT services can read, write and delete les and folders that are stored in
the cloud.
Error Handling: because applications and services are not operated locally or in
the company-internal infrastructure, it is important that they can report errors.
To integrate the error handling with support the Logistics Mall provides an
interface to the ticketing system. Using this interface an IT service can report
errors using a Web service or sending a simple e-mail to the ticketing system. In
both cases error handling is integrated with the ticketing workflow.
User and role management: a key requirement for the provisioning of services
and applications from the Logistics Mall is that after a one-time authentication a
user can call all its associated services and applications without repeated login
procedure (Single Sign On). The Logistics Mall therefore provides an interface
that can be used by IT services to be query for user and role information.

4 Conclusion

The Logistics Mall platform aims at providing a market place for logistics services
and business processes together with a cloud-based access and execution envi-
ronment. The currently Logistics Mall platform that is accessible under www.
logistics-mall.com for the public supports the offering of IT services using the SaaS
delivery model.
In the next step this platform will be extended towards support for the design and
execution of workflows that combine logistics IT services and physical logistics
activities using the BPaaS delivery model. With regard to the process execution a
proof of concept implementation is being realized in form of a reference scenario.
This reference scenario includes the execution of different types of business pro-
cesses for warehousing within the intralogistics domain. The activities range from
support for the incoming goods, the storage and picking to the shipping process and
the outgoing goods. Part of the implementation is the integration of peripherals
typically used in logistics, such as label printers and handheld scanners. The
modeling and design methodology using the LPD is also demonstrated and vali-
dated by the reference scenario. The results obtained so far suggest that the pro-
posed methodologies a viable, meaning that the proposed concept is suitable to be
used in industrial applications.

References

1. Doll, A., Friebel, D., Rckriegel, M., Schwarzmller, C.: Global logistics markets. http://
www.rolandberger.at/media/pdf/Roland_Berger_Studie_Global_Logistics_Markets_n_
20140820.pdf (2014). Accessed 24 Nov 2014
2. Forrester Research: Sizing the clouda BT futures report. Understanding and quantifying the
future of cloud computing. http://www.forrester.com/Sizing+The+Cloud/fulltext/-/E-RES58161
(2011). Accessed 08 April 2013
Logistics MallA Cloud Platform for Logistics 27

3. Mell, P., Grance, T.: The NIST denition of cloud computing. Working paper National
Institute of Standards and Technology. http://www.nist.gov/itl/cloud/upload/cloud-def-v15.pdf
(2009). Accessed 08 April 2013
4. Amazon Elastic Compute Cloud (Amazon EC2). http://aws.amazon.com/ec2/ (2013).
Accessed 08 April 2013
5. Microsoft Windows Azure. http://www.microsoft.com/windowsazure (2013). Accessed 08
April 2013
6. Salesforce CRM. http://www.salesforce.com (2013). Accessed 08 April 2013
7. Jeffrey, K., Neidecker-Lutz, B. (eds.): The future of cloud computingopportunities for
European cloud computing beyond 2010. Expert group report, European Commission, DG
INFSO (2010)
8. ten Hompel, M., Schmidt, T.: Warehouse Management: Automation and Organisation of
Warehouse and Order Picking Systems. Springer, New York (2007)
9. Meinhardt, M.B., Rahn, J.: Empirical qualitative analysis of the current market situation in the
context of cloud computing for logistics. In: ten Hompel, M., Rehof, J., Wolf, O. (eds.) Cloud
Computing for Logistics, Lecture Notes in Logistics, Springer (2015)
10. Schuldt, A., Hribernik, K.A., Gehrke, J.D., Thoben, K.D., Herzog, O.: Cloud computing for
autonomous control in logistics. In: Fhnrich, K.P., Franczyk, B. (eds.) 40th Annual
Conference of the German Society for Computer Science (GI 2010), LNI 175, Gesellschaft fr
Informatik, pp. 305310. Leipzig, Germany, 27 Sept-1 Oct (2010)
11. Kaisler, S., Money, W.H.: Service migration in a cloud architecture. In: 44th Hawaii
International Conference on System Sciences (2011)
12. Goyal, P., Mikkilineni, R.: Policy-based event-driven services-oriented architecture for cloud
services operation and management. In: IEEE International Conference on Cloud Computing
(2011)
13. Scholz-Reiter, B., Rippel, D., Sowade, S.: Limitations in modeling autonomous logistic
processeschallenges and solutions in business process modeling. In: IEEE International
Symposium on Assembly and Manufacturing (ISAM), pp. 16, 2527 (2011)
14. Bentounsi, M., Benbernou, S., Atallah, M.J.: Privacy-preserving business process outsourcing.
In: IEEE 19th International Conference on Web Services (2012)
15. Fraunhofer Innovation Cluster Cloud computing for logistics. http://www.ccl.fraunhofer.de/
en.html (2013). Accessed 08 April 2013
16. Bhmer, M., Schmidt, M., Weienberg, N.: Seamless interoperability in logistics: narrowing
the business-it gap by logistics business objects. In: ten Hompel, M., Rehof, J., Wolf, O. (eds.)
Cloud Computing for Logistics, Lecture Notes in Logistics, Springer (2015)
17. Holtkamp, B.: The logistics mallan it-architecture for logistisc-as-a-product. In: ten Hompel,
M., Rehof, J., Wolf, O. (eds.) Cloud Computing for Logistics, Lecture Notes in Logistics,
Springer (2015)
18. Bochon, I., Ivens, V., Nagel, R.: Challenges of cloud business process management. In: ten
Hompel, M., Rehof, J., Wolf, O. (eds.) Cloud Computing for Logistics, Lecture Notes in
Logistics, Springer (2015)
19. Eggemann, J., Leveling, J., Wei, N.: Business apps meet the challenge of covering
continually changing logistics requirements. In: ten Hompel, M., Rehof, J., Wolf, O. (eds.)
Cloud Computing for Logistics, Lecture Notes in Logistics, Springer (2015)
Empirical Qualitative Analysis
of the Current Cloud Computing
Market for Logistics

Maren-Bianca Wolf and Jonas Rahn

Abstract To determine the opinions, needs, and requirements of the market actors
in the cloud computing market for logistics, Fraunhofer IML conducted an
empirical market analysis with users and vendors of logistics and IT. Qualitative
interviews were used to examine closely the potential of this concept and any
barriers to it. The results of this analysis allow conclusions to be drawn about the
future development in the logistics and IT industry.

1 Introduction

Cloud computing is now a reality. After the initial furore and cautious hesitation,
the technology was able to hold its ground in the market and is now widely in use.
Various sectors of the industry have recognized the advantage of distributed
computing power and have put it to use in their respective areas of business. The
numbers are clear: The rapidly increasing budgets of the past few years give us a
sense of the size of the fundamental change in technology that it has brought with it
cloud computing has shaped and changed the IT market permanently in the same
way that a disruptive innovation would.1 Extremely volatile business areas with
strongly fluctuating requirements such as logistics benet greatly from the flexible
distribution of resources in the cloud.

1
Prser [21].

M.-B. Wolf (&)
J. Rahn

Fraunhofer Institute for Material Flow and Logistics, Dortmund, Germany
e-mail: [email protected]
J. Rahn
e-mail: [email protected]

Springer International Publishing Switzerland 2015 29

M. ten Hompel et al. (eds.), Cloud Computing for Logistics,
Lecture Notes in Logistics, DOI 10.1007/978-3-319-13404-8_3
30 M.-B. Wolf and J. Rahn

2 The New Standard

In many places the use of cloud computing is now viewed as a matter of course and
no further discussion is needed. For a few years this topic was still considered to be
a future trend and although it was the cause of heated discussions it was almost
never viewed as relevant for the policy of a company. This has clearly changed:
According to the Federal Association for Information Technology, Telecommuni-
cations, and New Media (BITKOM); cloud computing has been selected as the hot
topic for the information and telecommunications industry in 2013 for the fourth
time.2 The market research company Forrester Research predicted that for 2013
almost half of all North American and European companies will have a budget for
investing in private clouds and many software development managers will plan to
use cloud applications.3

3 Current Situation

According to estimates by the Experton Group, German companies will put about
5 % of their IT costs into the cloud in 2013.4 This means an increase in investment
of circa 52 % over 2012: The costs for cloud services, cloud integration, and
consulting as well as cloud technology in the business cloud will total approxi-
mately 4.6 billion euro in 2013. A growth of around 50 % has been forecasted for
2014. This is a total investment of 6.9 billion euro.5
According to other estimates, the cloud computing market in Germany will grow
by 47 % in 2013 (business and consumer cloud together), like it did the previous
year, and reach a volume of 7.8 billion euro (2012: 5.3 billion. euro). A sustained
growth of between 30 and 40 % is also forecasted for the following year.6 The
cloud market in Germany is expected to reach a sales volume of approximately
20.1 billion euro in 2016. This is according to a report by the high-tech association
BITKOM based on current forecasts by the Experton Group.
The European Commission released a statement announcing their new strategy
for using cloud computing to boost the EU GDP annually by 160 billion euro by
2020.7 According to BITKOM and the International Data Corporation (IDC), if the

2
In 2010, 45 % of the surveyed ICT companies indicated that cloud computing was the most
important IT trend. This was 62 % in 2011, 66 % in 2012, and 59 % in 2013. See the BITKOM
press releases from 13 January 2010, 18 January 2011, 18 January 2012, and 16 Januar 2013.
3
Staten [22].
4
Experton Group 2013: Janata and Velten: Current market gures for the German cloud com-
puting market5 % hit barriers [13].
5
Bayer [1].
6
BITKOM press release from 6 March 2013 (current forecasts of the Experton Group): Sales in
cloud computing rose to almost 8 billion euro [2].
7
Europische Kommission [5].
Empirical Qualitative Analysis of the Current Cloud Computing 31

legal framework is modernized and the broadband infrastructure is expanded, then

the annual boost to the GDP could be as much as 250 billion euro8 and 3.8 million
jobs could be created. Without political support the cloud computing boost to the
EU GDP could be as low as 88 billion euro.9

4 Making the Shift to Cloud Computing

The number of companies making the shift to cloud computing has changed greatly
in the past few years: Only 28 % of the companies surveyed in 2011 were open
minded about cloud computing and interested in it. This number rose to 35 % in
2012. But also, surprisingly, the number of sceptics also increased from 38 %
(2011) to 44 % (2012). The number of those who were undecided did go down,
which indicates that companies improved their information policies and their
uncertainty was dwindling. All of the discussions about the topic obviously helped
decision-makers make up their minds: The number of those who were undecided
shrunk from 33 to 20 %.10
The size of a company plays a role in how open they are to adapting the
technology: 55 % of companies with more than 2,000 employees were open to it
while smaller companies were clearly more reticent.11

5 Cloud Computing in Action

In 2011, 28 % of German companies were using at least one form of cloud com-
puting (another 22 % were planning to implement it or discussing its implemen-
tation)this number rose to 37 % in 2012 (planning rose to 29 %). Companies
from the information and telecommunications and nance sectors were the most
willing to adapt cloud computing. This group was followed closely by the trans-
portation and logistics, chemical and pharmaceutical, automotive engineering,
mechanical and plant engineering, and commerce sectors.
Some of the reasons for the positive attitudes of many companies towards cloud
computing are lower IT administration costs, shorter implementation times for new
applications and solutions, rapid scalability of IT services, improved mobility and
geographically distributed access to IT resources, lower IT costs, increased orga-
nizational flexibility, improved performance and availability, and a higher capacity
for innovation. The opportunity to tap into new customer groups and save costs

8
Kempf [14].
9
International Data Corporation (IDC): Quantitative Estimates of the Demand for Cloud Com-
puting in Europe and the Likely Barriers to Up-take, p. 9 [12].
10
KPMG AG, BITKOM & PAC: Cloud-Monitor 2013, p. 7 [17].
11
Ibid., p. 8.
32 M.-B. Wolf and J. Rahn

(hardware and personnel) also plays an important role. The main concerns are no
longer with data protection but with the loss of data. Other risks that have been
identied are a diminishing IT know-how and an unclear legal situation.12

6 Public Perception

Cloud computing still has a perception problem: According to a recent study from
CSC, a global provider of IT services, 70 % of Germans are worried about their
cloud data. They are primarily worried about their personal data in other countries
and about the loss of control that comes with it. They also do not trust that the cloud
service provider will perform the statutory delete requests reliably. The majority of
the Germans who were surveyed (86 %) want proof of data protection and technical
security from the cloud service provider. 82 % want the certication of the safety
standards to be required by law.13
According to BITKOM and in spite of all of these concerns, every second
Germany saves their photos, music, contacts, and calendars in the cloud.14 When
looking at only the 1429 old, the average cloud usage rate goes up to 97 %.15

7 Public Clouds Versus Public Clouds

Eighty-three percent of users of private cloud services rated their experiences as

positive in 2012, 25 % more than 2011. The rating for public cloud services
dropped a bit in that same period: there were only 74 % satised users in 2012
instead of 81 %.
Thirty-four percent of Germany companies used private cloud solutions in 2012
(an increase of 7 % from the year before, 29 % were planning to implement one or
discussing it) while 10 % decided to use a public cloud (an increase of 4 % from the
year before, another 11 % were in the planning phase). Two thirds of the public
cloud users were also using a private cloud. This means that many companies are
using public cloud applications to supplement their private cloud. More than three
quarters of all cloud users rated their experiences as positive.16

12
Ibid., p. 4.
13
CSC: IT-Sicherheit [4].
14
Matz [19].
15
Lber [18].
16
KPMG AG, BITKOM & PAC: Cloud-Monitor 2013, p. 4 [17].
Empirical Qualitative Analysis of the Current Cloud Computing 33

8 Current Difculties

8.1 Insufcient Utilization of Resources

The benets of cloud computing are still not fully being utilized. One of the reasons
for this is that there is still an inefcient usage rate of the cloud storage system:
While the ideal usage rate of stationary data carriers is circa 50 %, only 17 % of the
cloud storage is being used worldwide. The usage rate is a little bit higher in
Germany at 26 %. However, if only small and medium-sized companies are con-
sidered then this falls to 7 %. Small and medium-sized companies are not using
more than 90 % of their availableand paid forcloud storage. It is no surprise
that this is not helping to reduce costs.17

8.2 SLAs and Uniform Standards

Service Level Agreements (SLA) can help ensure that misunderstandings are
avoided by specifying the scope and quality of the service in detail. However, both
parties need the required know-how to create the SLA. The terms and conditions of
the SLA should dene the expectations between the two parties regarding avail-
ability of the services, the permissible failure rate, and the response time in case of
failures. The SLA also has to dene the consequences for non-adherence with the
stipulated terms and conditions as well as the exclusions and limitations. It also has
to contain the terms for terminating the SLA and the exact guidelines for returning
and deleting the data (for example, the contract ends or one of the companies goes
bankrupt).18

8.3 Data Protection and Safety

Although the cloud seems like it is ideal for the IT-intensive branches of the
industry because of its scalability many of the players are showing restraint. Out-
standing issues and unresolved preconceptions are still keeping some companies
back from outsourcing applications to the cloud. In reality, many of the misgivings
are unfounded: Cloud computing can actually help improve data protection, data
availability, and performance.19

17
Bckle [3].
18
Kilz [15].
19
KPMG AG, BITKOM & PAC: Cloud-Monitor 2012, p. 17 [16].
34 M.-B. Wolf and J. Rahn

Even in the PricewaterhouseCoopers survey Global State of Information

Security Survey 2012 the majority of the companies surveyed (54 %) indicated
that they improved their information security through the application of cloud
services.20

9 Users of Cloud Computing

The number of cloud users has increased: In 2011, 28 % of companies were

benetting from the cloud while this number rose to 37 % in 2012. The number of
companies planning to move to the cloud or at least discussing it climbed to 22 % in
2011 and to 29 % in 2012.21 According to the IDC study Cloud Computing in
Germany 2012Evolution of the Revolution, cloud computing was an ofcial
component of the IT strategy for 56 % of the companies surveyed and an ofcial
component of the business strategy for 34 % of the companies surveyed and this
was by the middle of 2012.22 The cloud seems to have established itself in medium-
sized companies as well: According to the IT Cloud Index Q4/2012, one quarter
of all medium-size German companies were using Saas, laaS, and/or PaaS solutions
by the end of 2012.
New trends such as big data, mobile applications, social business, and social
collaboration encourage the use of cloud computing because they require global
data availability as well as flexible computing and storage capacities.23
Small and medium-size companies benet the most from the cloud: They are the
ones who need custom costeffective solutions but most companies do not have
the budget to invest in numerous applications. With the advent of pay-per-use
solutions, small and medium-sized companies now have access to solutions they
could not afford before and they can now rent high-quality products without paying
high licensing fees.
It really seems like the only way for small and medium-sized companies to stay
competitive is for them to take the step into the cloud. Small companies just do not
have the powerful and scalable IT infrastructures that they need to manage the
constantly increasing volumes of data.24

20
PricewaterhouseCoopers (PwC): Key ndings from the Global State of Information Security
Survey 2012, p. 24 [20].
21
KPMG AG, BITKOM & PAC: Cloud-Monitor 2013, p. 10 [17].
22
International Data Corporation (IDC): [11].
23
Hofmann: Study conducted in Germany. IDC: Die Verzahnung von Big Data und Cloud
Computing wird enger und ECIN: Cloud Computing 2013Steigende Investitionen, Wach-
stumtreiber Social Collaboration und neue Sourcing Optionen [8].
24
Techconsult GmbH; Hewlett-Packard GmbH [23].
Empirical Qualitative Analysis of the Current Cloud Computing 35

10 Cloud Computing for Logistics

Several different surveys on cloud computing in the logistics sector have been
conducted in the past few months and published as studies. One of them was an
online survey conducted by the software provider INFORM GmbH which showed
that 68.3 % of the surveyed companies are ready right now to use cloud computing
for logistics tasksonly 12.7 % have actually done it. The reasons for this are a
lack of familiarity with the topic (29.5 %) and the security concerns mentioned by
almost half of the surveyed companies. The possibility of having to rely on an
external service provider was a barrier to using cloud technology for 13 % of the
surveyed companies. The lack of industry-specic solutions was an obstacle for
another 5 %. There seems to be a wide range of reasons. Flexible access (38 %),
reduction in operating costs (25 %), faster implementation times for business
processes (18 %), platform independence (12 %), and access to IT resources that
would not be possible without cloud computing (7 %) were identied as the ben-
ets. According to the respondents, cloud computing solutions can be used for the
communication between vendors and customers, controlling suppliers, and man-
aging supply chain events.25
The current trend in the cloud is focussed almost exclusively on the use of a
private cloud in the logistics industry. Only 6 % of the surveyed companies indi-
cated in Cloud Monitor 2013 that they use a public cloud solution. The logistics
industry and the retail and wholesale industry are bringing up the rear in terms of
public cloud usagewith the logistics industry slightly out in front.26
Even the ndings of the study on the Benets of Cloud Computing for
Logistics that the Institute for Cloud Computing (IFCC) conducted on behalf of
Axit AG conrmed the current trend: 60 % of the respondents were convinced that
the future belongs to the cloud.27 Logistics service providers are particularly open
minded about the technology: almost 63 % of them are interested in the cloud (more
than a third of them already have a cloud strategy). This is further proof that
discussions about the cloud help to increase the trust in the fact that the cloud is
here to stayreadiness grows with experience. As mentioned earlier, openness also
increases with the size of the company: Companies with more than 1,000
employees show an above average interest in cloud computing.28
Small and medium-sized companies in particular often do not have enough IT
expertise and capacity or the necessary capital to bridge the gap between the

25
INFORM GmbH: Die Logistik und die CloudVision oder Wirklichkeit? INFORMs trend
barometer shows that despite security concerns it is medium-sized companies in particular who are
open to the idea of using logistics IT services from the cloud [9].
26
KPMG AG, BITKOM & PAC: Cloud Monitor 2013, p. 26 [17].
27
Institute for Cloud Computing (IFCC); Axit AG: Studie zum Nutzen von Cloud Computing
fr die Logistik , p. 6 [10].
28
Institute for Cloud Computing (IFCC); Axit AG: Studie zum Nutzen von Cloud Computing
fr die Logistik , p. 15 [10].
36 M.-B. Wolf and J. Rahn

requirements and the status quo in logistics IT. Time is often the factor that prevents
companies from developing flexible and innovative logistics solutions in a number
of respects. The time it takes to implement the logistics systems of today is quite
long compared to the length of time the system is used. As a result, business models
and strategies in trade and commerce currently have a typical lifecycle of 23 years.
An average implementation time of 610 months is shown when adequate (logistic)
business processes are modelled. A new and more powerful IT solution is also
required to provide the requested logistics services and this extends the imple-
mentation time by 1224 months on average. The total usage time for the system
(circa 35 years) is clearly too short for a company from a strategic investment
point of view.
Pressure is also growing on the logistics industry because of some changes in
legislation. In addition to the often cited toll charges, the new Working Hours Act
and the necessity for electronic customs clearances both mean that new IT services
will need to be used, for example for controlling driving times and rest periods. The
same is true for the introduction of tracking and tracing for the continuous moni-
toring of goods.
It seems that the challenges placed on logistics services by both the industry and
technology represent an exemplary area of application for Software as a Service
solutions and cloud computing, which are viewed as the new paradigm for the
provision of IT services and are considered to be a growth market.29

11 New Cloud Computing Concepts for Logistics: The

Fraunhofer Innovation Cluster Cloud Computing
for Logistics

The Fraunhofer Innovation Cluster Cloud Computing for Logistics, which was
started by the Fraunhofer Institute for Material Flow and Logistics IML and the
Fraunhofer Institute for Software and Systems Engineering ISST, combines the
design and structure of logistics and information technology services in one com-
mon design. The planned research and development work cumulated in the
Logistics Mallas a central marketplace for everything from individual logistics IT
functionality up to complete process chains, which can be offered as a product. This
makes it possible for a company to put together the logistics services they need
when they need them and operate them in the Logistics Mall. The opportunity to
swap out an individual logistics IT service with the service offered by an alternative
vendor at runtime does not only result in new business models but it also avoids
dependencies on a service provider.
The Internet is both a technological platform and a signpost of the Logistics Mall
and cloud computing is the future of the Internet. The main benets of cloud

29
Hamburger [7].
Empirical Qualitative Analysis of the Current Cloud Computing 37

computing are going to be summarized here one more time. Cloud computing min-
imizes the costs and the barriers for starting the shift toward web portals. The reason
for this is that the computing power is provided by the provider so no large investment
in expensive infrastructure is needed. The scalability is another benet because the
computing power is adapted to the number of users. Cloud computing is the tech-
nology that is the safest bet for the future. One cloud consists of hundreds of thou-
sands of servers that are replaced with newer models and always kept up to date. This
means that the technology will never get old. Cloud computing also offers a high
degree of availability, flexibility, and sustainability. Companies can reduce their
ecological footprint because cloud computing helps them use their resources and
systems more optimally and efciently. The hottest topic is still the question of
guaranteeing data protection in the cloud. Fraunhofer IML is paving a path here for
the users and vendors in the Logistics Mall by offering a certication program for both
vendors and products to ensure compatibility with the Logistics Mall.
The Logistics Mall uses Web 2.0 technology and was developed for the logistics
of the future. With Web services as the basis, information and services within the
Logistics Mall cannot only be requested but also new high value logistics process
can be orchestrated. Like in an online store, the (virtual) user clicks on the desired
services and thereby uses a global, open software architecture. The applications are
hosted on the servers of the mall operators or the connected systems of the vendor.
The customer only pays for what they use. The operator of the Logistics Mall
handles the billing.

12 Results of the Empirical Analysis

Cloud computing will be indispensable to all industries in the future and the out-
sourcing of computer capacity to the cloud will be an essential component of every
business model. The topic is still controversial today. Many market participants are
worried about the security of cloud solutions and are not quite ready to make the jump.
The Fraunhofer Institute for Material Flow and Logistics IML examined the
maturity of the logistics market for these solutions in a study. They asked users and
vendors of logistics software about their attitude towards cloud computing tech-
nologies in a series of discussions.
Fraunhofer IML conducted the market analysis Cloud Computing for Logis-
tics in 2011 and 2013 to identify and analyse the opinions, needs, and require-
ments of the users and vendors in the logistics industry in terms of both cloud
computing and the Logistics Mall model. The project team conducted qualitative
interviews and used the results of those interviews as the basis for examining the
potential and barriers of this concept and determining which requirements needed to
be implemented in the Logistics Mall. By examining the acceptance and readiness
of those surveyed, they could draw conclusions about the attitudes of the logistics
experts in the industry. To perform the analysis, they compared reference values
from 2011 with the new values from 2013.
38 M.-B. Wolf and J. Rahn

The study participants from the users group were from small and large com-
panies from three branches of business: logistics service providers, industry, and
wholesaler/retailer.
The study participants from the vendors group were from small and large pro-
viders of logistics solutions.
The results showed that the spirit of innovation of the users group is split right
down the middle, which validated the theoretical assumptions that were made
before the project began: It is still too early for every second participant to use a
technology like the Logistics Mall. When asked more concretely 56 % of the
participants from the users group would be willing to use logistics software in the
cloudif the conditions were right in their own company. This value has actually
decreased by 8 % compared to the study conducted in 2011.
The interviews with the study participants from the vendors group in 2013
revealed that the innovative spirit of the vendors group is slightly less than it was in
2011. Half of the participants (49 %) are willing to sell their products or services
through an external online store, which is decline of 9 % from 2011. A total of 67 %
(70 % in 2011) would be willing to operate their applications in the cloud. In
contrast to 2011, 10 % of the study participants in the vendors group are already in
the concrete planning phase for using cloud architecture (Figs. 1, 2).
Can you imagine running your application in a cloud environment?

Fig. 1 The level of acceptance for selling products or services through an external online store
(vendors)
Empirical Qualitative Analysis of the Current Cloud Computing 39

Fig. 2 The level of acceptance for operating applications in the cloud (vendors)

The agreement with these two fundamental Logistics Mall ideas diverges sig-
nicantly with the size of the company. Small and medium-sized companies were
more willing to change their views than larger companies. When the Logistics Mall
was described in more detail, 58 % (75 %in 2011) of the vendors surveyed indi-
cated that they would use it. Sixty-seven percent (70 % in 2011) of the vendors said
that they were sure that their customers would be willing to use cloud computing in
logistics. A big change in the answer to the question about the change in the
relationship with the customer was seen when the 2013 data was compared to the
2011. Fifty-six percent (39 % in 2011) accepted the fact that as the vendor they
were no longer the main contact point for the user but that the operator of the
Logistics Mall held that role instead.

13 Opportunities and Risks

In addition to publicly known risks of performance and security (63 %), the users
also named the following as areas of concern: failure of the cloud server (27 %),
excessive complexity (23 %), problems communicating with existing systems
(20 %), the insolvency of the vendor (14 %), the possible development of price
dumping (7 %), the possible formation of a cartel (6 %), and a lack of differentiation
from competitors (4 %).
40 M.-B. Wolf and J. Rahn

According to the vendors, the main advantages for them and their customers are
the opportunity to tap into new customer groups and the potential cost savings in
personnel and hardware. They also hope that a synergy effect will result from the
integration of their applications with the complementary applications of other
vendors. This would mean that they would be able to offer customized solutions to
their customers. They still have concerns about data protection and data security but
these are more psychological barriers for the customers than problems with the
technology (Fig. 3).
The participants from the users group were asked to rate several aspects of cloud
computing in logistics on a scale from 1 (very important) to 6 (not important at all)
as shown in Fig. 4. The respondents rated safe encryption of the data (1.5) and a
guarantee that the applications run quickly and smoothly (1.6) as the two most
important aspects. Twenty-four hour availability, customer support (2.2), and

Fig. 3 Success factors for the user

Empirical Qualitative Analysis of the Current Cloud Computing 41

Fig. 4 Success factors for the vendor

access to high-quality software that the company could not normally afford (2.2)
were rated as important. Smooth communication between the system interfaces and
the ability to build a process chain using services from different vendors were both
rated as important with a score of 2.3.
The main requirements of the vendor for the operator of the Logistics Mall are a
free trial of the Logistics Mall (1.8), targeted advertisements for the vendors
products by the operator (2.1), vendor access to a Content Management System
(CMS) to describe their own products and services (2.1), provision of the hardware
infrastructure (2.2), and vendor access to a CRM system to monitor the activities of
their own customers on the Logistics Mall website (2.3).
Twenty-six percent of users (22 % in 2011) are willing to enter into a contract
with the operator of the Logistics Mall instead of the vendor. There is demand for
the billing model (static or dynamic) and the contract periods and all possible
congurations will soon be offered.
In 2011 37 % of the surveyed companies used WMS, whereas in 2013 only
32 % of the responding companies reported to operate such software. The utili-
zation of TMS, however, has increased slightly: in 2011 30 % of the evaluated
42 M.-B. Wolf and J. Rahn

companies used this kind of software and in 2013 already 32 % of the companies
did that could provide information about their software application. Regarding
active ERP systems an even bigger discrepancy is identied: whereas in 2011 one
third of the surveyed users conrmed the utilization, in 2013 already 46 % of the
respondents applied an ERP system. The DMS use also increased from 27 % in
2011 to 34 % in 2013.
The reason that some participants in the users group have never used logistics
software is that their company is too small to have special systems to support their
logistics processes and, as a result, they did not have the budget for it. Twenty-ve
percent (30 % in 2011) of those who have used logistics software indicated that
they have thought about changing vendors. The willingness to change is low
because they think that the costs associated with making this change are too high
and cannot even be justied by the improvements in performance and functionality
they would get by changing vendors. Often they can get their current vendor to
make some improvements by complaining so it seems to them that it is no longer
necessary to make a change.

14 Results

The results clearly show the spirit of innovation of German managers: with an
acceptance rate of 56 % this value has sunk by 12 % compared to 2011. Despite this
decline, this number still shows that more than half of the respondents are ready and
willing to rent logistics software from the cloud today. They answered the question
Can you imagine using logistics software that is not running locally on computers
at your company but instead on servers in the Internet? with yes. Those who
answered no revealed in the qualitative interviews that one of the reasons was that
they did not feel familiar enough with the topic. It is important to note that the
concept of the Logistics Mall had not been explained to them yet when this question
was asked. The goal of this question was to determine their attitude towards the idea
of cloud computing in logistics. Some of those who said no to cloud computing in
the survey explained that it was because they thought the technology was not
developed enough for them to make a shift in their IT strategy.

15 Conclusion and Outlook

The increasingly homogenous range of services and products offered in the logistics
market are making it harder and harder for a customer to select the right vendor.
From the vendors perspective (for small and medium-sized companies), it can be
difcult to stay competitive because the main focus is always on the price and
certain well-known companies have put their stamp on the market over the years.
Large companies also face the challenge of staying competitive because customers
Empirical Qualitative Analysis of the Current Cloud Computing 43

keep asking for more and more features. Customers want flexible and tailor-made
solutions and they want them on demand. This demand from customers has
driven the need in the future for a suitable marketplace where companies can offer
their logistics solutions. Just like a shopping mall, a B2B customer wants to be able
to select the products and services they want and combine them with others. To be
able to continue to increase customer satisfaction and also satisfy all demands, two
disciplines need to mesh and become better integrated in the futurelogistics and
marketing.
Customers do not look for goods or services per se, they look for solutions.30

The quote clearly shows that customers are thinking bigger than just services and
products. They want a solution for their problem. They want added value at the end
of a process. It all comes back to individuality. The Logistics Mall as a solution for
providing custom service offerings fosters the interactive character of the solution
and leads to increased customer satisfaction. This new approach, one of viewing
and using the Logistics Mall not just as a platform for products and services but as a
solutions platform, needs to be lived. In the future this will mean for the customer
that none of their problems will go unsolved.
The results of the market analysis make it clear that the time for cloud computing
and, in particular, cloud computing in logistics is here. The logistics market is
already moving in that direction.
The business culture will change over time. The focus will be on custom
solutions. The network concept that the Logistics Mall represents leaves a lot of
leeway. Each companyno matter how big or smallhas the chance to be a part of
this development and help shape it: The experts are already saying that the Logistics
Mall will become the largest logistics platform in the industry.

References

1. Bayer, M.: Die Cloud wird gesellschaftsfhig, in: COMPUTERWOCHE, online, 6 Februar
2013: http://www.computerwoche.de/a/die-cloud-wird-gesellschaftsfaehig,2532192
2. BITKOM Presseinformation vom 6. Mrz 2013 (aktuelle Prognosen der Experton Group):
Umsatz mit Cloud Computing steigt auf fast 8 Milliarden Euro, online: http://www.bitkom.
org/de/markt_statistik/64086_75301.aspx
3. Bckle, R.: Symantec-Studie. Versteckte Kosten in der Cloud, in: ChannelPartner, online, 5.
Februar 2013: http://www.channelpartner.de/2601143
4. CSC: IT-Sicherheit: 70 % der Deutschen haben Angst um ihre Cloud-Daten, online, 6.
Dezember 2012: http://www.csc.com/de/press_releases/93261-it_sicherheit_70_%25_der_
deutschen_haben_angst_um_ihre_cloud_daten
5. Europische Kommission: Mehr Schwung fr das Cloud Computing, online, 27. September
2012: http://ec.europa.eu/news/science/120927_de.htm

30
Grnroos: Service Management and Marketing. A Customer Relationship Management
Approach, p. 4 [6].
44 M.-B. Wolf and J. Rahn

6. Grnroos, C.: Service Management and Marketing: A Customer Relationship Management

Approach, Wiley, Chichester (2000)
7. Hamburger Abendblatt: Microsoft und Telekom erwarten Durchbruch fr Cloud Computing,
online, 07.10.2010:http://www.abendblatt.de/wirtschaft/article1654862/Microsoft-und-Telekom-
erwarten-Durchbruch-fuer-Cloud-Computing.html
8. Hofmann, K.: In Deutschland durchgefhrte Studie. IDC: Die Verzahnung von Big Data und
Cloud Computing wird enger , in: IT-Business, online, 1. Februar 2013: http://www.it-
business.de/index.cfm?pid=4840&pk=393096&print=tr
9. INFORM GmbH: Die Logistik und die CloudVision oder Wirklichkeit? INFORM-
Trendbarometer zeigt: Trotz Sicherheitsbedenken ist vor allem der Mittelstand offen fr
Logistik-IT-Services aus der Cloud, online, 18. Januar 2013: http://www.inform-software.de/
produkte/invent-xpert/news/article/die-logistik-und-die-cloud-vision-oder-wirklichkeit/
10. Institut fr Cloud Computing (IFCC); Axit AG: Studie zum Nutzen von Cloud Computing
fr die Logistik , online, Juli 2012: http://www.axit.de/cloudlogistik/
11. International Data Corporation (IDC): Cloud Computing in Deutschland 2012Deployment-
Modelle und Management, Integration, Security und Compliance im Fokus, online, Juli 2012:
http://www.idc.de/consulting/mc_cloud2012_reg.jsp
12. International Data Corporation (IDC): Quantitative Estimates of the Demand for Cloud
Computing in Europe and the Likely Barriers to Up-take, online, Juli 2012: http://ec.europa.
eu/information_society/activities/cloudcomputing/docs/quantitative_estimates.pdf
13. Janata, S., und Velten, C. (Experton Group): Aktuelle Marktzahlen zum deutschen Cloud-
Computing-Markt5-%-Hrde erreicht, online, 30. Januar 2013: http://www.experton-group.
de/research/ict-news-dach/news/article/aktuelle-marktzahlen-zum-deutschen-cloud-
computing-markt-5-%25-huerde-erreicht.html
14. Kempf, D.: Vortrag im Rahmen der Pressekonferenz zum Cloud Monitor 2013, online, 21. Februar
2013: http://www.bitkom.org/les/documents/BITKOM_Vortrag_PK_Cloud_Monitor_21_02_
2013_nal_web.pdf
15. Kilz, S.: Wasserdichte Vertrge und SLAs aushandeln, in: Computerwoche, online, 27. Februar
2013: http://www.computerwoche.de/a/wasserdichte-vertraege-und-slas-aushandeln,2533212
16. KPMG AG, BITKOM & PAC: Cloud-Monitor 2012, Hannover, online, 6. Mrz 2012: http://
www.kpmg.de/docs/cloud-monitor-20120529.pdf
17. KPMG AG, BITKOM & PAC: Cloud-Monitor 2013, Hannover, online, 6. Mrz 2013: http://
www.bitkom.org/les/documents/BITKOM_KPMG_PAC_Studie_Cloud_Monitor_2013.pdf
18. Lber, K.: Die Clouddes einen Freund, des anderen Feind, in: Die Welt, online, 27. Februar
2013: http://www.welt.de/sonderthemen/cebit-2013/article113928252/Die-Cloud-des-einen-
Freund-des-anderen-Feind.html
19. Matz, A.: Die Wolke im Griff, in: Hamburger Abendblatt, online, 5. April 2013: http://www.
abendblatt.de/ratgeber/article115026738/Die-Wolke-im-Griff.html
20. PricewaterhouseCoopers (PwC): Key ndings from The Global State of Information Security
Survey 2012, online, September 2011: http://www.pwchk.com/webmedia/doc/634653330
562192188_rcs_info_security_2012.pdf
21. Prser, S.: Die CloudIn aller Munde, aber noch lange nicht im Handel, auf: heise resale,
online, 16. November 2011: http://www.heise.de/resale/artikel/Die-Cloud-In-aller-Munde-
aber-noch-lange-nicht-im-Handel-1331924.html
22. Staten, J.: 2013 Cloud Predictions: Well nally get real about Cloud, online, 3. Dezember
2012: http://blogs.forrester.com/james_staten/12-12-03-2013_cloud_predictions_well_nally_
get_real_about_cloud
23. Techconsult GmbH; Hewlett-Packard GmbH: IT-Cloud-Index, Q4/2012, online, April 2013:
http://www.techconsult.de/index.php/43-uncategorised/231-clouindex1-13
The Logistics MallAn IT-Architecture
for Logistics-as-a-Product

Bernhard Holtkamp

Abstract The Logistics Mall extends the cloud service model from Software-as-a-
Service (SaaS) to Business Process-as-a-Service (BPaaS). Its architecture, hence,
spans from a web-based public B2B marketplace for simple apps, traditional
applications and even logistics process templates to customer-specic cloud based
execution environments for instantiated process models and therein included
applications. For communication with the mall outside world a specic gateway is
provided that controls inbound and outbound message flow.

Keywords Logistics mall Cloud computing
Software-as-a-service (SaaS)

Business process-as-a-service (BPaaS)

1 Introduction

Currently the time for implementing new logistics systems is long compared with
their usage time. In trade and industry business models and strategies have a life
cycle of 23 years. The development of new logistics processes often takes
610 months. If new IT systems are needed to support these processes, imple-
mentation time increases to one or even up to 2 years. This is too long and too
expensive for strategic investments. The use of logistics specic cloud services is
therefore considered as a viable solution.
Hence, the Logistics Mall provides virtualization of IT resources and business
processes to relieve especially small and mid-size production and logistics com-
panies from investments in the development of individual software solutions
including expensive software licenses and from operating costs. Thus, these com-
panies can focus on their core business.

B. Holtkamp (&)
Fraunhofer Institute for Software and Systems Engineering ISST,
Emil-Figge-Str. 91, 44227 Dortmund, Germany
e-mail: [email protected]

Springer International Publishing Switzerland 2015 45

M. ten Hompel et al. (eds.), Cloud Computing for Logistics,
Lecture Notes in Logistics, DOI 10.1007/978-3-319-13404-8_4
46 B. Holtkamp

Logata is a key partner of the Fraunhofer innovation cluster Cloud Computing

for Logistics that develops the Logistics Mall. Logata is the provider and operator
of the Logistics Mall as a public cloud solution. Low margins and growing com-
petition in the logistics market on the one hand and Logatas position as an SME on
the other hand require a low cost solution. To avoid expensive license fees for
software components the Logistics Mall is implemented as an aggregation of open
source software solutions.
In the following we have a closer look at the Logistics Mall architecture. Fol-
lowing the TOGAF Architecture Development Model [1] we start with the Business
Architecture. The architecture vision of the Logistics Mall is already described in
detail in [2]. In a next step we discuss some design decisions that are relevant for
the understanding of the technical architecture that is presented in subsequent
sections of this document. We present the Logistics Mall version one that applies
the Software-as-a-Service paradigm rst and continue with enhancements of this
version towards a higher level of integration of applications within the Logistics
Mall and with external applications up to the level of Business Process-as-a-Service
in the nal version of the Logistics Mall. We close with a look to the current status
of usage and with an outlook to further developments.

2 Business Architecture

The description of the business architecture focuses on three aspects: business

strategy, organization and key business processes to be supported by the Logistics
Mall.

2.1 Business Strategy

Logata is the Logistics Mall provider and thus the face to the customer. Logata
aims at providing a broad high-quality offering. This means, like in a shopping
mall, a balanced portfolio of products and competing suppliers. To guarantee high
quality suppliers and their products are certied. The focus is on suppliers of
logistics IT applications and of business process support. Hence, the Logistics Mall
provides a rich portfolio of services for suppliers to appear as an attractive sales and
distribution channel. That includes CRM functionality for offerings and follow-ups
of such offers, Infrastructure-as-a-Service, help desk and 1st level support and last
but not least billing support for suppliers towards their customers.
In addition to the role as Logistics Mall provider Logata also intends to act as a
supplier to offer their own solutions in the mall. The target group is formed by
SMEs from the logistics and production domains with annual revenues in the range
of 70 Mio. Euros or an IT budget of about 1.4 Mio. Euros. As pricing model a
The Logistics MallAn IT-Architecture for Logistics-as-a-Product 47

monthly flat rate is envisaged. The goal is to provide customers with IT solutions
that reduce their IT costs by 30 %.
In a later stage the target group will be extended towards global players.

2.2 Organization

Logata is the provider of the Logistics Mall as a public cloud solution. In that role
Logata is in charge of acquiring suppliers of logistics IT applications and to perform
marketing and PR activities to create awareness for the Logistics Mall in the cus-
tomer target groups.
As the face to the customer Logata also provides a helpdesk and 1st level
support to end users, i.e. registered customers that have booked services in the
Logistics Mall. For suppliers Logata acts as a clearinghouse, doing billing and
invoicing towards the suppliers customers in the mall and billing towards the
suppliers for the use of Logistics Mall services.
Suppliers offer IT applications as Software-as-a-Service solutions and provide
2nd level support for their applications.
In the rst phase, Logata is also the technical operator of the Logistics Mall.

2.3 Business Processes

When looking at business processes to be supported by the Logistics Mall we have

to differ between the following stakeholder roles: Logistics Mall provider, Logistics
Mall operator, potential mall customer, logistics IT supplier and end user. On the
end user side we can further differentiate between administrator, normal user and
manager as a privileged user.
Potential customers of the Logistics Mall can navigate through the public part of
the Logistics Mall homepage on the Web and can register for becoming a mall
customer.
Registered end users can book applications, services and processes in the
Logistics Mall. They can use the booked functions, receive usage reports and billing
information. In case of problems the users can contact the helpdesk or can submit
trouble tickets.
Registered suppliers can apply for the inclusion of their applications and services
in the Logistics Mall portfolio. If a product is accepted the supplier gets the means
to manage the product in the Logistics Mall catalogue. Suppliers also receive usage
reports and billing information. They have to provide 2nd level support for their
products. In case of problems they can also contact the helpdesk or submit trouble
tickets.
The Logistics Mall provider is in charge of marketing and PR processes to attract
suppliers and users for the offered solutions. The Logistics Mall provider checks
48 B. Holtkamp

registrations and endorses or rejects applications. The provider also performs the
administration of the Logistics Mall, including customer management, billing and
reporting. Logistics Mall customers are suppliers of IT applications as well as users
who book such services and use them in the Logistics Mall cloud. The Logistics
Mall provider also runs the helpdesk and provides 1st level support to its customers.
The Logistics Mall operator runs the Logistics Mall infrastructure including
monitoring and backup of the Logistics Mall. The operator also sets up usage
environments for new customers and integrates new applications in the mall
environment.

3 Design and Implementation Decisions

Security and maintenance aspects were driving forces of the following design and
implementation decisions. Security is a big issue anyway regarding the acceptance
of cloud computing, especially in Germany. Motivating companies to use business
critical applications in the cloud implies the adoption of adequate security mea-
sures. Providing a company exclusive environment is an adequate measure.
Vendor lock-in is another issue that prevents companies from migrating their
applications into the cloud of a particular cloud service provider. An exclusive
environment per customer makes it easier for the cloud service provider to return
customer data to their customer and to remove the customer from the cloud without
leaving security and privacy relevant traces.

3.1 Design Decisions

A key decision for the Logistics Mall is the separation of the commercial part from
the usage part. That leads to a Logistics Mall Marketplace (MMP) with the portfolio
of offered applications, services and processes. Customers can register here and
registered customers get access to their private part of the marketplace to book mall
products, submit calls for tenders to suppliers or read their usage reports or billing
information. Suppliers get services to manage their products in the mall catalogue,
i.e. edit product descriptions or their company prole.
Users, having booked applications in the mall, get access to their services via a
Customer Access Framework (CAF). Each customer company gets its own CAF to
ease the management of service level agreements as well as backup/restore and the
management of maintenance windows.
To reduce central administration efforts, self-administration of customers is
applied to user management, to communication management via the Logistics Mall
Gateway and to peripheral devices at customer premises that are used by applica-
tions in the mall.
The Logistics MallAn IT-Architecture for Logistics-as-a-Product 49

A single sign on service is provided that covers the MMP, a customer CAF and
the applications integrated in the CAF.

3.2 Implementation Decisions

As mentioned before a general implementation decision with regard to license fees

was the use of Open Source software as far as possible. An exemption was made for
the virtualization layer. Regarding the experience with VMware and the product
quality at the time the decision was made it was decided to use VMware vSphere
for the virtualization layer.
After the evaluation of various portal software solutions [3] Liferay was taken as
a basis for portal development. For security and maintenance reasons it was decided
to have different portal instances for MMP and CAF. Moreover, each customer
company gets its own CAF portal on its own virtual machine.
Components that are shared between MMP and CAFs (e.g. user management or
basic services (reporting, trouble ticketing, and accounting)) are implemented in a
Logistics Mall Infrastructure layer (LMI).
As long as applications are not multi-tenant ready each customer gets their
booked application instance on its own virtual machine.

4 Logistics Mall (Stage 1)

To enable a fast time-to-market the Logistics Mall is developed in steps. In the rst
step the Logistics mall offers logistics IT applications applying the Software-as-a-
service paradigm (Fig. 1).
Customers book one or more applications in the Logistics Mall Marketplace
(MMP) and get access to the booked applications through their Customer Access
Framework (CAF). In the following we have a brief look at the logical architecture

SaaS
Customer
Provider Customer
SaaS Customer
Provider
SaaS
Logistics Customer
Provider
SaaS Mall Customer
Provider Customer
SaaS
Public Cloud Customer
Provider Customer
SaaS
Provider

Fig. 1 Logistics Mall as public cloud SaaS platform

50 B. Holtkamp

of the Logistics Mall, followed by a closer look at the technical architecture.

Finally, MMP and CAF are discussed in more detail.

4.1 Logical Architecture

The logical architecture of the Logistics Mall can be seen from different perspec-
tives. From the viewpoint of the stakeholders the Logistics Mall looks as depicted
in the following gure. A purchaser of a logistics service provider or of a pro-
duction company basically sees the MMP. He uses the MMP for collecting
information about the Logistics Mall portfolio for registration as a customer. After
registration he uses the MMP for booking applications. The management of the
customer company might also use the MMP for accessing usage reports and billing
information. Customer administrators use the customer exclusive CAF for man-
aging users, their roles and access rights, and for managing peripheral devices that
must be available to booked applications. Users of the customer company log into
their CAF to access the booked applications.
In case of problems with the Logistics Mall or with booked applications any user
can submit trouble tickets via MMP or CAF.
The 1st level support of the Logistics Mall provider uses the Web interface of the
integrated trouble ticketing system to handle trouble tickets from their customers
(Fig. 2).
From a technical perspective the Logistics Mall looks as depicted below. MMP
and CAF portals reside on top of the Logistics Mall Infrastructure (LMI). LMI and
the portals use components from the basic software platform underneath.

Fig. 2 Logistics Mall (stage 1) stakeholder view

The Logistics MallAn IT-Architecture for Logistics-as-a-Product 51

Fig. 3 Logistics Mall logical architecture from a technical viewpoint

The bottom layer provides the virtual machines for the upper layers. It is
implemented by VMware vSphere 4 Essentials Plus (Fig. 3).

4.2 Technical Architecture

On a gross level MMP and CAFs have the same architecture. The frontend part
consists of the presentation layer and a service integration component. The backend
part is composed of the service layer and the persistence layer. The components are
discussed in more detail below (Fig. 4).
Figure 5 illustrates the implementation of CAF components using basic open
source software platform components. My WMS and Puzzle represent applications
that are integrated in the CAF. The black arrows show the control flow from a user
initiated request in the browser down to persistence layer components.

4.3 Presentation Layer

The Logistics Mall presentation layer provides different interfaces for the different
user roles identied in Sect. 2.3 as well as interfaces to functionality independent of
a users role. Furthermore, interfaces are different for MMP and CAFs.
52 B. Holtkamp

Fig. 4 Logistics Mall gross architecture

Fig. 5 Logistics Mall (stage 1) technical architecture

The MMP presentation layer provides stakeholder specic interfaces. An

unregistered visitor gets access to public information like e.g. application portfolio
or supplier catalogue and to the registration service. When a buyer (from a regis-
tered customer company) logs in, the MMP provides access to the list of booked
products, to the related list of suppliers and to the list of pending orders. When a
supplier logs in, the MMP provides service interfaces to product management,
offers management, customer contact management, product registration and sup-
plier activities. The mall operator is provided with content management function-
ality for news and help, customer registration management services, system
administration services, services related to operator activities and with product
management services (Fig. 6).
In addition to stakeholder specic services the MMP provides common services
for single sign on, customer and user administration, reporting and reporting
The Logistics MallAn IT-Architecture for Logistics-as-a-Product 53

Fig. 6 MMP presentation layer components

Fig. 7 CAF presentation layer components

administration, accounting, news access, trouble ticketing, searching, document

management and CAF access.
The CAF presentation layer provides service interfaces for the user roles of a
customer. A regular user gets access to the applications he has a role in. Managers
get access to reporting and accounting services and to reporting administration.
Administrators are provided with service interfaces for application management,
document management, reporting management, and for the management of
peripheral devices. All customer user roles get interfaces to common services like
single sign on, user administration, helpdesk and document management (Fig. 7).
54 B. Holtkamp

Fig. 8 Service layer components

4.4 Service Layer

The Logistics Mall service layer provides web services related to product and
application management, related to customer and user management and for com-
mon services (accounting, document management, reporting, trouble ticketing). The
Web Content Management services support MMP content management. The
Peripheral Device Management services support the management of peripheral
devices on a customers premise that are connected to a booked application in the
Logistics Mall cloud (Fig. 8).

4.5 Persistence Layer

The persistence layer manages the Logistics Mall non-volatile data. Key compo-
nents are the product catalog, tenant registry, user registry, roles registry and MMP
content. The product catalog is divided into two parts: the product records contain
information about the product types offered in the MMP, the product instance
The Logistics MallAn IT-Architecture for Logistics-as-a-Product 55

Fig. 9 Persistence layer

components

records contain information about the instances of product types that are set up
when booked by a customer. The tenant registry manages company related infor-
mation of suppliers and end user companies that have registered. The roles registry
keeps the roles that are supported by the products in the MMP product catalog. The
user registry keeps the user account information of the Logistics Mall tenants
(Fig. 9).

4.6 Implementation

The Logistics Mall is implemented by means of open source software. Key com-
ponents are already mentioned in Sect. 4.2 when describing the technical archi-
tecture. Figure 10 shows the implementation layers with their related technologies.
The Logistics Mall components are deployed on different virtual machines as
illustrated by Fig. 11.
In our development and test environment the Logistics Mall virtual machines run
on the physical infrastructure depicted in Fig. 12.
56 B. Holtkamp

Fig. 10 Logistics Mall implementation layers

Fig. 11 Logistics Mall deployment schema

The Logistics MallAn IT-Architecture for Logistics-as-a-Product 57

Fig. 12 Logistics Mall physical infrastructure (development and test environment)

5 Logistics Mall (Stage 2 + 3)

The Logistics Mall in stage 2 extends the SaaS model supported in stage 1 by
providing application integration support for applications in a CAF as well as for
applications in the mall with applications outside. Integration support in a CAF is
provided by the Logistics Mall Bus and by the Business Object Repository. Inte-
gration of CAF internal applications with applications outside of the Logistics Mall
is supported by means of the Logistics Mall Gateway. The Logistics Mall in stage 2
and 3 supports not only traditional applications but apps as well. From the Logistics
Mall point of view an app executes IT-based activities as part of logistics processes.
Apps can be combined to complex workflows. They can be interactive or fully
automated. The functionality ranges between simple services in the sense of SOA
and complex applications. Apps communicate by exchanging business objects via
the Logistics Mall Bus within a CAF or via the Logistics Mall Gateway with
external parties. They can also share objects within the Business Object Repository.
More details on Logistics Mall Apps are given in [4].
The Logistics Mall Bus enables an asynchronous exchange of Logistics Mall
Business Objects between connected applications within a CAF. The bus provides a
web service interface that accepts a business object from a sending application. The
bus buffers a business object and passes it to a receiving application via web service
request.
The Business Object Repository stores instances of Logistics Mall Business Object
types. The repository can store the objects that are used within an application but can
also act as a shared data source for multiple applications. A detailed discussion of
Logistics Mall Business Objects and their repository can be found in [5] (Fig. 13).
58 B. Holtkamp

Fig. 13 Logistics Mall (stage 2 + 3) logical view

The Logistics Mall in stage 3 also provides for logistics business process
management from process design to process execution. The web-based Logistics
Process Designer supports Logistics Mall compliant design of logistics processes
by combining apps from the mall to executable workflows. These processes can
then be deployed on a process engine that forms an integral part of a CAF in stage
3. For more details on process management in the Logistics Mall we refer to [6].

5.1 Logistics Mall Bus

The Logistics Mall Bus represents the internal communication infrastructure of a

CAF. It provides web service interfaces for the exchange of Logistics Mall Business
Objects between connected applications. Business objects are transferred as Busi-
ness Object Documents according to OAGIS [7] in XML format.
The Logistics Mall Bus provides the TransferBOD webservice server that is
called by an application or by to pass a business object to the bus. The bus transfers
a business object into the queue of the receiving application. For each connected
application the bus maintains such a queue for buffering BODs until they can be
passed to the receiving application. The ReceiveBOD webservice client transfers
the BODs from an applications queue via webservice call to the application
connector (see Fig. 14). The bus log records all transfers so that the flow of business
object can be traced if necessary.
Figure 15 shows the Logistics Mall in stage 2 from a technical perspective.
The Logistics Mall Bus is implemented by means of the open source talend
enterprise service bus [8]. For the buffering of messages Apache JMS queues are
used [9].
The Logistics MallAn IT-Architecture for Logistics-as-a-Product 59

Fig. 14 Transfer of BODs between applications in a CAF

Fig. 15 Logistics Mall (stage 2) technical view

5.2 Logistics Mall Gateway

The Logistics Mall Gateway is the interface of the Logistics Mall to the outside
world. The gateway coordinates and controls the communication between appli-
cations in the mall and their outside peers. Information exchange is only allowed
60 B. Holtkamp

Fig. 16 Logistics Mall gateway architecture

between a product instance in a CAF and registered external parties. Partners and
connections are managed in corresponding gateway directories. Each mall customer
is enabled to manage communication connections between their booked applica-
tions and their external partners. For the information exchange different channels
can be used. Currently, web services and ftp is supported, other channels can be
added when needed. As exchange format Logistics Mall Business Objects are
preferred, other electronic document types can be used as well, but in combination
with a converter application in the mall.
The gateway restricts the communication to authorized parties only. For the
exchange of Logistics Mall Business Objects the gateway also limits the transfer to
object types specied for a dened connection with a registered partner, e.g. partner
A may only send purchase orders per web service to the order management system
in a CAF if the partner is listed in the gateway directory and the connection is
dened there accordingly. For each business object the gateway also checks its
structural correctness (Fig. 16).
The Logistics Mall Gateway is implemented by means of the 4D rapid appli-
cation development environment [10]. The Logistics Mall runs a single gateway
server and each CAF gets its own LMG console that is connected with the mall
clients part of the gateway server.
The Logistics MallAn IT-Architecture for Logistics-as-a-Product 61

6 Conclusion

The Logistics Mall architecture covers the demands of three different stakeholders:
end users, suppliers of logistics apps, and mall operator.
End users have a high demand for security besides strong demands for flexibility
and availability. That is taken into account by providing an end user with a com-
pany exclusive usage environment for apps and processes. Asynchronous com-
munication between apps in the mall and external parties via dened connections
and for well-dened business objects increase the security level signicantly. Only
valid objects from identied sources can pass the gateway, using secured com-
munication channels. The exclusive use of the mall infrastructure assures high
availability and performance as side effects from other mall clients are excluded. It
also eases the exiting from the mall as client data can be easily returned to the mall
customer and can be easily removed in the mall as well.
Suppliers of apps are provided with a dened test environment for checking the
mall compliance of their apps. The communication interfaces provided by the
Logistics Mall ease the communication of business objects.
The Logistics Mall basic services for accounting, reporting and trouble ticketing
also support the supplier business so that he can focus on app development and
product management as his core competencies.
The mall operator benets from the self-administration functionalities of the
Logistics Mall Gateway and of the CAF. The client exclusive usage of the mall
infrastructure eases the control and management of service level agreements.
While the Logistics Mall is operational in stage 1 for pilot customers, the next
development steps still have to prove its viability in practical use. This holds in
particular for stage 3 with the composition and deployment of logistics processes.

References

1. The Open Group: TOGAFTM Version 9. Van Haren Publishing, Zaltbommel (2009)
2. Daniluk, D., Holtkamp, B.: Logistics MallA Cloud Platform for Logistics. In: Cloud
Computing for Logistics, Lecture Notes Logistics, Springer (2015)
3. Weienberg, N., Stemmer, M.: Modern IT-Platforms for Business Process Management and
PortalesComparative Evaluation of the Tool Suites from IBM, IDS Scheer & SAP, and
intalio & liferay. Fraunhofer ISST, Dortmund (2009) (in German)
4. Wei, N., Leveling, J.: Business Apps Meet the Challenge of Covering Continually Changing
Logistics Requirements. In: Cloud Computing for Logistics, Lecture Notes Logistics, Springer
(2015)
5. Bhmer, M., Schmidt, M., Weienberg, N.: Seamless Interoperability in Logistics: Narrowing
the Business-IT Gap by Logistics Business Objects. In: Cloud Computing for Logistics,
Lecture Notes Logistics, Springer (2015)
6. Bochon, I., Ivens, V., Nagel, R.: Challenges of Cloud Business Process Management. In:
Cloud Computing for Logistics, Lecture Notes Logistics, Springer (2015)
62 B. Holtkamp

7. OAGi: Business Object Document (BOD) Message Architecture for OAGIS Release 9.+.
OAGi White Paper Document #20110408V1.0. http://www.oagi.org/oagi/downloads/
ResourceDownloads/2011_0408_BOD_Message_Architecture_V9x.pdf (2013). Accessed 12
April 2013
8. Talend ESB: http://www.talend.com/products/esb (2013). Accessed 12 April 2013
9. Java Messaging Service: http://www.oracle.com/technetwork/java/jms/index.html (2013).
Accessed 12 April 2013
10. 4D: http://www.4d.com/ (2013). Accessed 12 April 2013
Business Apps Meet the Challenge
of Covering Continually Changing
Logistics Requirements

Julian Eggemann, Jens Leveling and Norbert Wei

Abstract This paper describes a business app approach that is used within the
Logistics Mall to develop and to execute autonomous IT services. These services
are called Apps because they focus on a single business function. In this paper the
three different kinds of business apps and the architectural components are
described based on continually changing logistics requirements. The key aspect of
this article is the Workbasket component for apps, including a graphical user
interface to manage a task list.

Keywords Logistics Mall Business apps SOA Services Logistics Mall Apps

App workbasket

1 Introduction

Today a typical logistics process runs in a heterogeneous IT environment. Such

systems are mostly standard software like SAP, Oracle, or more specialized
logistics software systems like Warehouse Management Systems (WMS) or Pro-
duction Planning Systems (PPS). All these systems do not entirely meet the
requirements of logistics customers regarding short contract duration, pay-per-use
accounting and the provisioning of individual IT services. The majority of the
features of standard software is rarely used or not used at all, but has to be paid for,
due to a monolithic architecture. Additionally, the unneeded functions are often

J. Eggemann
J. Leveling (&)
N. Wei

Fraunhofer Institute for Material Flow and Logistics IML,
Joseph-Von-Fraunhofer-Str. 2-4, 44227 Dortmund, Germany
e-mail: [email protected]
J. Eggemann
e-mail: [email protected]
N. Wei
e-mail: [email protected]

Springer International Publishing Switzerland 2015 63

M. ten Hompel et al. (eds.), Cloud Computing for Logistics,
Lecture Notes in Logistics, DOI 10.1007/978-3-319-13404-8_5
64 J. Eggemann et al.

running in the background and consuming resources. In the end, this leads to more
unnecessary costs. Finally, customizing software to adjust according to new or
changed requirements involves high efforts, costs and risks. Nevertheless, the
solutions flexibility for further adoptions is not increased [1].
Apps are the answer to these requirements. The Business App Approach pre-
sented in this chapter is based on Service-Orientation as described by [2] and [3]
combined with the specic logistics requirements as shown in the next sub-chapter.
In order to compose complex processes using existing applications there is a need
for modularization, not only as an architectural aspect of a monolithic application
but for the whole application. Business Apps are a ne grained approach to reflect
the needs of individual processes in the logistics IT. The key feature of an App is to
deliver only a single business function.
Finally the composition of these ne grained Apps to a process enables a run-
ning business, supported by IT solutions that are covering all individual require-
ments. Furthermore, the composition of Business Apps allows the adjustment of
modelled processes due to continually changing logistics requirements without the
ballast of unneeded service functions as pointed out by [3].

2 Logistics Requirements

Logistics often focuses on minimizing the costs over several processes. In this
branch process optimization and cost reducing methods like two-stage order-
picking are common. This method separates the order-picking-process into two
different sub-processes. During the rst sub-process all articles for all orders are
collected from the stock. In the second step, all articles need to be separated and
attached to the given orders.
The Logistics branch requires highly available and easy to use software solutions
to resolve complex process chains. The additional requirement is an open interface
structure to allow partner companies to exchange data. Furthermore, the logistics
companies depend on efcient and stable workflows which need to support by
flexible software solutions to resolve the new workflows or process changes.
In general, within a cloud process three different kinds of tasks can be distin-
guished. The rst is a converter service for the integration of external systems. All
external systems have their own data structure and are only able to communicate
with general data specication within the logistics domain, like EDIFACT [4]. This
data has to be converted between the given specication and the Logistics Mall
internal business objects (BO) in both ways of communication. The BOs are
logistics domain specic objects, like a picking list or a handling unit (HU). The BOs
are transport objects, which allow the transport of data in-between the running
application or between different apps. For each required data specication a single
converter app is required related to the one app one function principle. Integration of
external data is needed to migrate existing data into the cloud environment and to
connect cloud services with systems hosted on premise. These services are executed
Business Apps Meet the Challenge of Covering 65

autonomously. The second service kind is for executing autonomous business

functions like calculations and does not require user interactions. Those functions are
common in the logistics domain. Finally the third service kind is for providing a user
interface. For each of these kinds a special type of app is dened. The three apps,
converter-, service-, and end-user-apps, are described in detail in chapter Empirical
Qualitative Analysis of the Current Cloud Computing Market for Logistics.
For all these kinds a dedicated Business App has to be developed. The one App
one business function principle as mentioned in chapter Logistics Software
Systems and Functions: An Overview of ERP, WMS, TMS and SCM Systems
addresses the continually changing requirements of the logistics domain. If the
customers requirements have changed, a process could be easily adjusted by
removing, replacing or adding different apps. Another benet of this granular
approach is that a customer has to pay only for functionality that is really needed
and executed. Furthermore, this rule may leads to small services, which are easier to
monitor and to manage as complex services with multiple functions. Additionally,
these small services can be exchanged with other services in a running process.
Due to the Service-oriented approach Business Apps are loosely coupled and
should have no knowledge about other services or environment specics. They have
to provide an interface that is able to accept Business Object Documents (BOD) and
to send BODs to the Process Engines Message Broker. A BOD can be described as
an envelope that contains a list of BO instances of one type. The BOD itself contains
information about the origin of the BOD. BODs that are exchanged with the Process
Engine are only containing BO instance IDs. The BO instance details have to be read
from the BO Instance Repository by the App themselves. The Broker is dened as
central bus infrastructure for communicating between Logistics Mall environment1
essentials. Furthermore, Apps have to request the required BO instances, which are
presented by their IDs within the BOD, from the BO Instance Repository. These
three ways of communication must be provided and implemented by each Business
App that should be executed within the Logistics Mall Cloud environment.
The Business App architecture should also be technology neutral to enable the
adaption of already existing IT services for the Logistics Mall. Those services are
based on several technologies. The lowest standardized common denominator of in
the Logistics Mall is used by the environment. The communication infrastructure
should be designed with widely accepted standards as dened by [2]. This enables a
wide range of technologies a Business App could be developed with.

3 Logistics Mall Specics Requirements

In this chapter requirements of other Logistics Mall components are described. The
focus is here mainly on the process engine that interacts as mentioned before with
the apps.

1
More details in chapter The Logistics MalAn IT-Architecture for Logistics-as-a-Product.
66 J. Eggemann et al.

The process engine is not able to differentiate between App types. Additionally,
the Business Object Instance Repository is not able to lock BO instances due to its
client wide usage. Further, the repository does not have any knowledge about Apps
or other clients, like the process engine. It is just a system wide data storage
approach. To consider this issue, End-user Apps have to provide a mechanism to
store incoming BOs until a user completely processed it. This mechanism has to
avoid errors due to parallel processing of one BO. A BO has to be locked directly, if
a user started to process it.

4 Business App Architecture Approach

This chapter brings the Logistics requirements and the Cloud environment together
and describes how our Logistics Mall Apps cover these requirements by an
application view. For a better overview, this chapter names requirements and
presents the cloud environment. Later the Logistics Mall App structure and the
different App architecture layers are described. The mentioned requirements can be
realized by cloud solutions, such as the Logistics Mall. The following paragraph
describes the cloud environment used by the Logistics Mall Apps.
The Logistics Mall offers a cloud environment to fulll the above mentioned
logistics requirements. The Logistics Mall represents a marketplace for logistics
software and services which can be accessed over the web-front-end designed by the
cloud operator. The designed architecture based on the Software as a Service (SaaS)
service model. In this model software systems are executed in a cloud environment
and accessed through a Web Browser as described by [5]. Furthermore, the cloud
infrastructure is managed by the cloud operator. Each customer can book the required
software apps. Before use, the booking needs to be approved and provisioned.
Afterwards, the customer is able to simply use the booked applications over the web-
browser. The advantage of the mentioned purchase process is based on the flexible
use of the booked software solutions running on cloud solutions, such as the Logistics
Mall. Based on the flexible use, the apps are offered on a pay-per-use billing. Next to
the pay-per-use billing option, other product subscriptions can be offered, such as a
full servicepackage. The way of billing is arranged by the application provider. A
main advantage of the Logistics Mall is based on the creation of individual logistics
processes. This is done by the Logistics Process Designer (LPD).2
Moreover, the service providers can offer their services to customers by adding
their Apps to the Logistics Mall Marketplace. Before the applications can be used
within cloud environment, the service providers need to adapt their software
applications to the standards, employed by the Logistics Mall. In addition to that,
the App has to be adapted to the cloud computing solution.

2
More details are available at chapter Challenges of Cloud Business Process Management.
Business Apps Meet the Challenge of Covering 67

After introducing the Logistics Mall and the basic idea of Apps being used in the
Logistics Mall, the following paragraph focuses on the structure inside the cloud
infrastructure. The major difference between local software solutions and cloud-
based software solution is the different delivery of the functions. While local
software needs to be installed on local hardware, the cloud-based solution is pro-
vided as a service on the external environment. An environment for Business
Process as a Service (BPaaS) solution requires a more complex infrastructure to
support a high number of customers.
A booked app is provided as an instance to every customer. Each instance could
be used within multiple processes. Multiple instances of one app will only be
created, if an instance is highly used and more memory and processing power are
required as the app or the app execution environment can manage itself. Addi-
tionally, the Logistics Mall delivers for each version of an app a single instance, if
an App is needed in different versions. The companys data, as well as the booked
instances, are stored in the cloud-environment. Due to privacy and security reasons,
the data is separated from other customers data by offering an individual platform
per Logistics Mall customer. The data is stored in a general database, called the BO-
Instance-Repository.3

4.1 App Instances

An App could be deployed on several Application servers. Each deployment of an

App within the Logistics Mall is named App Instance.

4.2 Layers

The following paragraphs describe the Logistics Mall App structure including the
different layers of the used structure. An App running on the Logistics Mall cloud
platform is divided into three layers as shown in Fig. 1. The layers are based on the
Model-View-Controller (MVC)Pattern as described by [6] extend by the
Workbasket that is described later on. The Model of each App are the used BOs.

4.2.1 Graphical User Interface (GUI)

The rst layer realizes the interaction with the clients running the application. This
is implemented by a web-UI to allow flexible use on most devices such as mobile

3
More details are available at chapter Seamless Interoperability in Logistics Narrowing the
Business-IT Gap by Logistics Business Objects.
68 J. Eggemann et al.

Fig. 1 Logistics Mall App

Layers

phones, tablets and computers. The user will be guided through most processes. The
dialogue usually begins with the start dialogue (start section) and nalizes the app-
process with the end dialogue.

4.2.2 Business Logic

The business logic interacts with the GUI and represents the algorithms and logic
behind each application. All functionality is run and organized by the business logic
while the transfer of data is structured by standardized objects, known as the
business objects (BOs).
Additionally, the business logic can retrieve business object data (processing
BOs) from the BO Instance Repository to achieve functional goals which cannot be
reached with the information given from the incoming BOs.

4.2.3 Workbasket

The Workbasket approach, as explained in this sub-chapter, addresses the

requirements for a mechanism to store all incoming messages from the process
engine and for locking these, if a user started the processing of a BO. The main duty
is to store all incoming Mini-BODs until a BO referenced by one BOD is com-
pletely processed. As shown in Fig. 2 there is one Workbasket instance for each
executed app instance that is shared through all user sessions of this app instance.
Business Apps Meet the Challenge of Covering 69

Fig. 2 Workbasket singleton

class diagram <<Singleton>>
Workbasket

-bodList : Map<String, BOD>

-lockedBOs : List<String>
+lockBO(sessionId , boId)
+unlockBO (sessionId, boId)
+unlockBOsBySessionId (sessionId )
+isLocked (boId) : boolean
+remove(boId) : boolean
+getBOIdList () : List <String>
+resolveBOIdtoBODId (boId) : UUID

The Workbasket contains also methods for locking and unlocking the BOs within
the received BODs. Every web user interface instance has its own unique id, which
is used to lock Mini-BODs, this disable the parallel processing of one BO instance
by two or more users.
The Workbasket is the part of the presented approach, which enables the split
and join of process instances. For example, an order (the input BO instance) with
four positions is split into two pickings lists (output BO) by a Business App. The
output of the App are two single BODs containing each one of the picking list BOs.
The next App instance within the process processes both BODs separately, but is
able to add parts of both pickings lists to one new delivery. All the splits are leading
to new process instance to handle the separate BODs.
A new entry to the Workbasket BOD-List is automatically added if the APP
instance RESTful interface is called by the Process Engine. The App approach is
designed for Java Enterprise Edition Version 6. The adding of a new BOD is
handled by the Context and Dependency Injection (CDI) event mechanism. That
allows a loose coupling between the RESTful interface and the business logic and
GUI of an App Instance Fig. 3.
After a BO instance is completely processed, it should be removed from the
Workbasket. Additionally the Process Engine will be informed by message about
the by instance process nalizing. Afterwards, the process engine is able to continue
the process instance by sending the results of the app instance to the next app along

<<RESTful>>
AppResource

+addMessage(BOD<Process, BO>)

Fig. 3 Workbasket interface

70 J. Eggemann et al.

the modeled process. If a user session is terminated due to an error or a session

timeout by an idle user, all locked BODs are unlocked automatically by the session
shutdown and other users are able to process the released BODs. The Workbasket
provides the required functionality for this.

5 Communication Within the Logistics Mall Infrastructure

The communication of an App instance with the Process engine and the BO-
Instance Repository is handled as shown in Fig. 4. The different Apps have to
contain a functionality to send processed BO information to the Process Engine.
Business Apps have to provide an interface, which is called by the process engine
to send new jobs to an App instance. The functionality and message characteristics
are described within this sub-chapter.
The communication within the cloud environment is implemented by a RESTful
API. A main advantage in using REST inside the Logistics Mall is the HTTP-based
structure, which allows an easy transfer of information throughout the whole cloud-
environment. Furthermore, REST is a lightweight and easy to use implementation
compared to other projects [7]. The Logistics Mall uses Business Objects (BOs) to
transfer information over REST, through the supported types XML and JSON.

Fig. 4 Architecture from App Instance perspective

Business Apps Meet the Challenge of Covering 71

More precisely, an app is transferring information based on a dened structure.

This structure is specied by BO-Denitions. A BO instance is transferred within a
Business Object Document (BOD). A BO Instance contains all necessary infor-
mation about a real domain specic object like a shipment or a handling unit. The
Business Apps receive minimized BODs from the Process Engine. These messages,
called Mini-BODs, contain only an ID of a BO instance. This is part of the
interface provided by the App. That means the App describes the BOD-type that
could be processed. All other types are not accepted.
The App has to load this instance. Afterwards, an App is able to process the
loaded BO for further operations within the given workflow.

5.1 App Characteristics

This chapter outlines the common structure of apps deployed in the Logistics Mall
infrastructure. The applications can be categorized in three different groups. Each of
the groups is presented below.

5.1.1 Converter Apps

The applications categorized in the converter app-group deploy the functionality

to translate the application-data of external Systems into the BO standard. In fact,
the application-data is converted to the equal Business Object used in the Logistics
Mall to deliver the given information in a valid format. Furthermore, the Converter
App also allows converting the BOs used in the Logistics Mall to the valid format
of the connected software solution. Every Converter App accepts exactly one
format. Communication with external systems is only possible through the
Logistics Mall Gateway.4 Additionally, converter Apps receive messages not solely
from this gateway, but from the Message Broker. All incoming external Messages
are sent to the Message Broker rst due to security considerations Fig. 5.

5.1.2 Service Apps

The second category of apps is known as Service Apps. The main purpose for
Service Apps is to deliver one business function to the Logistics Mall environment.
For example this could be a load space optimization algorithm. These apps do not
provide a user interface or contain a workbasket. The service delivered by a Service

4
See chapter The Logistics MalAn IT-Architecture for Logistics-as-a-Product for more
details.
72 J. Eggemann et al.

Fig. 5 Converter Apps

App will be automatically executed, if the provided RESTful interface is called by

the process engine.
The RESTful interface of a Service App has the same signature as the already
described workbasket interface, but no BO Ids are be administrated Fig. 6, 7.

5.1.3 User Apps

The last group of applications is summarized as User Apps. A User App describes
the interaction with different clients. While traditional software solutions can be
used over Graphical User Interfaces (GUI) to accomplish tasks, the Logistics Mall
User Apps offer a simple access to the user interface via a web-front-end. Over the
Web-front-end the user can simply run the functions due to his level of restrictions
(his role). Finally, the User Apps representing the only interaction between a user
and the cloud-computing service.

Fig. 6 Service Apps

Fig. 7 User App

Business Apps Meet the Challenge of Covering 73

6 App Life Cycle

The Business App Life Cycle is divided into ve steps. The following Fig. 8
describes the workflow within the App Life Cycle. The ve steps are detailed
within this chapter.

6.1 Development

The rst step is the development phase. An App Provider implements a service
following the guidelines described in chapter Empirical Qualitative Analysis of the
Current Cloud Computing Market for Logistics. The hard requirements are that
RESTful interfaces and service calls must be implemented. Additionally, the BO-
stack including BODs and Mini-BODS, of the Logistics Mall environment must
also be used for communication and the BO Instance Repository must be used for
storage of processed information and data shared by different apps of a process.
Furthermore, an end-user and the service App has to contain the workbasket
mechanism. Additionally, points are just suggestions to the provider, like the usage
of the Java enterprise stack. The developers are free to choose their own pro-
gramming language, but must make sure that their apps are executable within the
cloud environment. This is ensured and veried during the next phase of
the Logistics Mall App Life-Cycle. The development phase nishes with submitting
the created App and integrating it into the Logistics Mall Marketplace (MMP). For
the integration the apps description, its price model and the date of availability are
registered in the MMP. A Business App is only available until the specied date.
But rst of all the App is not visible or purchasable for any customer as long as the
Logistics Mall Verication has not been successfully completed.

Fig. 8 App Life-Cycle

Development

Undeployment/ Logistics Mall

Termination Verification

Execution Deployment
74 J. Eggemann et al.

Operator
IntegrationTests Stress Tests
Validation

Fig. 9 Verication phase

6.2 Logistics Mall Verication

All Business Apps have to be veried. Apps are tested to make sure that they are
covering the hard requirements. Also the tests verify a stable execution of Apps.
Additionally, apps are analyzed for malware5 detection. The main focus of this
phase is the stable execution within different processes. This is tested and validated
by stress tests and integration tests Fig. 9.
During the integration tests an app is deployed and executed within a test
environment. For the stress test an App instance is called sporadically with both
correct and erroneous messages. The erroneous messages are not compatible with
the RESTful interface of the App, they do not have the right format or they are not
even complete. This kind of stress test ensures a stable execution of any Business
App in the Logistics Mall environment. This phase is executed by the Logistics
Mall operator. If the test fails, the app provider has to do correct the occurred errors
and requirement lacks. If successfully veried the App is visible within the MMP
and entitled to be purchased and used for process modeling by any customer.

6.3 Deployment

Once a process is purchased by a customer, the operator starts the Apps deploy-
ment in case the app has not been instantiated before. Due to the successful veri-
cation this should be done without any errors.

6.4 Execution

The customer executes the new App instance within his processes. As mentioned in
chapter. Empirical Qualitative Analysis of the Current Cloud Computing Market
for Logistics end-user apps are executed by any user of the customer company.
Both Converter- and Service-Apps are executed automatically.

5
Malware are programs that are containing malicious operations.
Business Apps Meet the Challenge of Covering 75

6.5 Undeployment/Terminating

If an App Instance is not used within any process instance, the operator will stop this
instance gracefully. Afterwards, the App instance has to be marked as terminated
in the Logistics Mall Database. Once all instances of an App are terminated and the
App availability has exceeded, the App Status has to be set to expired. The App
provider is still able to add a new version of the App. Additionally, the availability
could be renewed.

7 Conclusion

The workbasket approach successfully bridges the gap between the needs of
logistics applications and the process-oriented view of the Process Engine. This
gives the opportunity to develop applications that can overcome the single process
instance focus of BPM with the side effect that processes will become more complex
due to changes of granularity. In our example of two-stage order-picking this means
a change from dealing with picking orders which are each owned by a single process
instance to the challenge to deal with picking order positions of multiple picking
orders in the rst stage. The second stage has to ensure the fulllment of the initial
order by joining the results of the rst stage and dividing the intermediate picking
positions according to the needs of the original orders. This change of granularity
will lead to additional process instances which have to be synchronized. Balancing
the aspect of increasing the complexity of the process model versus encapsulating
complex functionality in single apps will be a future task.
The presented architecture of the apps has proved to be satisfactory. Easy to
understand and maintain for the developer, flexible and lightweight but providing a
good framework for implementing many different apps. Granted that there are
enough apps to cover all needed processes the goal of providing the desired
functionality at lower costs and higher flexibility can be reached in the future.

References

1. Wolf, M.B., Rahn, J., ten Hompel, M.: Cloud computing fr logistik 2, fraunhofer verlag,
ISBN 978-3-8396-0558-5 (2011)
2. Papazoglou, M.: Web Services & SOAprinciples and technology, 2nd edn, isbn 978-0-273-
73216-7, pp. 778779 (2012)
3. Erl, T.: Service-oriented architectures, 2nd Edn, Prentice Hall, isbn 978-0-132-344-821, (2008)
4. EDIFACT Standards Overview Tutorial: Learn About Key e-Commerce Trends and
Technologies at Your Own Place, GXS, Online, http://www.gxs.co.uk/wp-content/uploads/
tutorial_edifact.pdf01 Aug 2014
76 J. Eggemann et al.

5. Holtkamp, B., Steinbu, S., Gsell, H., Lffeler, T., Springer, U.: Towards a logistics cloud,
presented at the 2010 sixth international conference on semantics, knowledge and grids, pp. 305
308 (2010)
6. Elias, S.J., Shahuddin, A.Z., Shaharudin, M.R.: Development of ISE club information system
(ISECIS): based on model-view-controller (MVC) pattern, LAP lambert academic publishing,
ISBN: 978-3-8443-2211-8 (2011)
7. Reitbauer, A., Novakovic, M.: RESTlos glcklich? in javamagazin, pp. 5658 (2009)
Seamless Interoperability in Logistics:
Narrowing the Business-IT Gap
by Logistics Business Objects

Martin Bhmer, Michael Schmidt and Norbert Weienberg

Abstract Logistics IT is shifting from large, monolithic software systems to

smaller, modular applications (apps) that can be flexibly composed to support
individual business processes. In order to anticipate extensive integration projects,
seamless interoperability is a crucial requirement. This paper outlines a standards-
based logistics Business Object model that narrows the business-IT gap and sim-
plies logistics IT services regarding development and integration.

Keywords Logistics Mall Business Objects model OAGIS Business-IT gap

Process modelling

1 Introduction

Logistics processes and supply chains are highly dynamic. They usually involve a
huge number of parties and rely on a variety of IT systems of different providers.
Establishing and maintaining these processes according to always changing
business requirements poses a big challenge on both the business level and the
technical level. The traditional separation of (and the gap between) business and IT
is an additional obstacle to overcome.

M. Bhmer (&)
M. Schmidt

Fraunhofer Institute for Material Flow and Logistics IML,
Joseph-Von-Fraunhofer-Str. 2-4, 44227 Dortmund, Germany
e-mail: [email protected]
M. Schmidt
e-mail: [email protected]
N. Weienberg
Fraunhofer Institute for Software and Systems Engineering ISST,
Emil-Figge-Str. 91, 44227 Dortmund, Germany
e-mail: [email protected]

Springer International Publishing Switzerland 2015 77

M. ten Hompel et al. (eds.), Cloud Computing for Logistics,
Lecture Notes in Logistics, DOI 10.1007/978-3-319-13404-8_6
78 M. Bhmer et al.

Flexibility and modularity of (physical) logistics services and IT systems have

become crucial aspects for rapid process adaption. Hence, extensive IT systems are
increasingly subdivided into smaller software applications (apps). When composed
to business processes, seamless interoperability of these apps is a requirement for
signicant competitive advantage. Considering the business-IT gap, standardisation
is necessary on both levels in order to narrow the gap and accelerate process
adaptations. This is what Business Objects are about.
The Fraunhofer Institute for Material Flow and Logistics IML and the Fraun-
hofer Institute for Software and Systems Engineering ISST have developed the
communication standard Business Objects for Logistics (BO4L [1]) in cooperation
with the Leading-Edge Cluster LogistikRuhr [2]. The main objective is to dene the
rst independent standard for the logistics domain, namely BO4L, which enables
logistical IT applications from different vendors to seamlessly interoperate (nar-
rowing the business-IT gap).
A Business Object (BO) represents an object relevant to a business domain, such
as a product, an order, or a handling unit. There may be a physical counterpart for a
BO, as for a product, but also non-physical objects like a pick list can be described.
A BO has a structured set of attributes that allows a comprehensive description of
the object for various applications.
Business Objects are prerequisite for the implementation of extensive, modular
logistics IT solutionsparticularly with regard to cloud concepts such as the
Logistics Mall [3, 4]. The BO4L standard is already in production use in a German
drugstore ordering system [5], presented at the German IT Gipfel 2012 [6]. In this
context, the Business Objects enable a continuous, standardised communication
between an online store, an enterprise resource planning system (ERP) and a
warehouse management system (WMS).
BO4L constitute a business-driven domain model and a technical standard for
interoperability at the same time. Thus, they do not solely target IT systems or
applications on a technical level, but also the communication of the involved parties
on a business or analyst level. Merging business and technical aspects, the BOs
enable domain experts and analysts to model, deploy and maintain business process
chains without signicant intervention by IT personnel. This closes the business-IT
gaps signicantly, and also closes some horizontal gaps, as detailed in this paper.
BO4L is not only a data model with different views, it is accompanied with
supporting software: a BO Instance Repository and a BO Type Repository, both
with Web browser support.
Information is the new oil [7] for the industry and this oil has to be stan-
dardized with respect to its types (the BO model) and its de-duplicated instances
(the BO Instance Repository, being a central data service). Hereby, development
and maintenance costs as well as time-to-market are minimized.
The paper is outlined as follows. Section 2 describes the main characteristics of
the Logistics Business Object Model BO4L, including its functional domain model
and several technical model representations, as well as procedure models for model
design and model governance. The issue of narrowing the business-IT gap by the
use of Business Objects and the closing of horizontal gaps is discussed in detail in
Seamless Interoperability in Logistics: Narrowing 79

Sect. 3. Since BO4L is more than merely a data model with different views, but also
supports persistence of instances and types, these supporting software components
for the BO model are explained in Sect. 4. Finally, background and related work are
presented. The paper concludes with a summary of all achievements provided by
BO4L.

2 The Logistics Business Object Model BO4L

We present the objectives for developing the Business Object model, the resulting
design and structure of its functional and technical layers, followed by corre-
sponding development methods and the models governance process.

2.1 Objectives

The objectives of a standard Business Objects model are on the one hand to
conciliate logistics and IT and on the other hand to mediate between different IT
systems (ERP, WMS, etc., see [8]) and parties. The former yields a common
understanding of Logistics and IT departments. The latter has two aspects: to
support xed communication of IT systems and to support a flexible orchestration
of IT systems using business process models by dened standard data objects for
logistics process models, as described in Sect. 3.4.
Hence, the business-IT gap must be narrowed vertically (Logistics-IT) and
horizontally (along the process chain). The vertical gap separates the business
process and data from their IT representations, while the horizontal gap comprises
an IT-to-IT gap and a Business-to-Business gap. This is illustrated by Fig. 1.
These objectives result in various requirements for the domain model and its
technical representations. A key requirement is the use of standards as described
below. Finally, the requirements are summarized by our denition of the term
Business Object.

2.1.1 Requirements for a Business Object Model

Targeting the vertical business-IT gap, the BO model would ideally merge the two
sides, so that there is single representation only. But in this case merging also implies
mixing of aspects specic to each of the sides and not helping the other side. Instead,
the model should consist of an abstract, domain-driven model (domain view) with
one or more technical implementations, each extending the abstract model by its
specic characteristics. The domain view and its technical representations (referred
to as technical view) must be coupled tightly to effectively bridge the business-IT
gap. Both views have to be extensible, versioned synchronously and controlled by a
80 M. Bhmer et al.

Fig. 1 Vertical and B2B gaps

horizontal business-IT gaps
Prepare Fulfill Bill
order order order
(Company A) (Company B) (Company A)

data
Business
IT GUI GUI
data
GUI
data
GUI GUI

WMS

ERP
IT2IT gaps

governance process, due to the complexity and dynamics of the (logistics) domain.
Different model versions must not interfere with each other in a process execution
environment when used simultaneously, e.g. by different processes.
The domain view abstracts from any technical details and can easily be understood
by domain experts and business users. It species relevant physical objects (e.g.,
products, palettes, inventory) and virtual or logical objects (e.g., pick lists, order lines,
storage areas) and their relations by reflecting the real world (domain).
The set of BOs must form a domain model enabling a consistent description of
comprehensive logistics processes. This includes dening (business) entities by
means of reusable, structured components, basic attributes (e.g. numbers, codes,
strings) and relationships to other entities and components. For ease of under-
standing these denitions should be provided in a graphical representation based on
a standard notation like UML. The domain model should be canonical (i.e., having
simplest archetypical structure) and based on or compatible to existing standards as
far as possible. Respecting established standards and extending them is the key for
broad acceptance. Communication using BOs may be legally binding, requiring the
specied entities to represent business transactions.
The technical representations should also be standards-based and enable
seamless integration on different IT layers for realising short time-to-market. The
ultimate goal is to support simple and flexible composition of apps from different
vendors. Thus, the technical representations of BOs are employed to specify the
apps interface proting from the semantically precise and vendor-independent
domain model. BOs are the key to app exchangeability, meaning an app can be
replaced by another app as long as the in- and outgoing BOs and the apps
semantics match. In combination with the domain model, the technical represen-
tations/views of BOs must be suited for creating exchangeable application building
blocks (apps). Finally, the technical views have to simplify app development in a
way that enables app developers to focus on their business logic.
The BO model is the key to avoid unwished dependencies. It does not only
decouple domain layers and technical layersat the same time it connects them by
Seamless Interoperability in Logistics: Narrowing 81

being a common language for both: The BO model decouples higher domain-
oriented architecture layers (especially business process models and their modeling
tool and execution engine) from lower technical layers (especially services and
applications). This increases the flexibility and agility of adapting process models
and of exchanging service implementations, since the BOs are not service-specic,
but provided similarly by any substitute service. This is explained in detail in [9],
pp 8286. For example, a one-step order-picking app can be replaced by a two-step
picking app without affecting the rest of the process.
The structure of BOs is the same for different use cases. However, different use
cases have different constraints: When using BOs for query by example (resulting
all objects that are similar to this sample) or for simple unmarshalling (e.g. con-
verting XML to Java), BO attributes must be allowed to be empty. However, when
communicating them to partners or when storing them, they have to fulll all
constraints dened by the domain. This also applies for using BOs with different
verbs or methods. The constraints a BO has to fulll also depend on its state: a BO
has a lifecycle, and in each state, other attributes of the BO are required, others are
optional. Therefore, the validity of BOs is a situation aspect and not an integral part
of a BO denition.

2.1.2 Benets of Using Standards

BO4L was intentionally designed to harmonize with existing standards. Open,

accepted standards result from the condensed experience and best practice of the
standard development team.
Using such standards contributes to the standardization of interfaces and prod-
ucts, as e.g. detailed in [10] and [11]: Unied information in the form of Business
Objects used across the entire value chain enables reuse and exchangeability of
(physical and/or business) services and software modules, ultimately cutting costs,
reducing risks and complexity as well as increasing data quality.
Providers (of services or software) can target larger markets, while users face
lesser risk of vendor lock-ins and prot from intensied competition among pro-
viders. Both providers and users gain investment security.
Moreover, lower transaction costs result, since standards accelerate business
processes, shorten negotiations, and simplify contracts. Standards for technical
interface parameters between subsystems enable an efcient construction of com-
plex systems and lower development costs, especially in cases involving a lot of
sub-systems, frequent changes and agile development.

2.1.3 Business ObjectA Denition

Business Objects specify real or virtual entities relevant to a particular domain by

structured sets of attributes, reusable sub-components and relationships to other
BOs. They serve as standardised means of communication for business processes
82 M. Bhmer et al.

and the IT systems these processes rely on. Thus, BOs are suited for specifying
interfaces on a business and on a technical level to facilitate Business-to-Business,
IT-to-IT and Business-to-IT integration. A BO has an abstract domain view and one
or more technical representations [12].
A BO model (in domain view) reflects a domain model enabling a consistent
description of comprehensive business processes in that domain. It abstracts from
all technical details not relevant to business users or domain experts. E.g., in a
business process model usually only the names and states of BOs and a few control
flow-relevant attributes will appear, as explained below.
A BO in technical view, i.e. a technical representation of a BO, is a software
artefact that implements the corresponding domain object with all relevant technical
details required for a particular purpose or technology. Thus, a BO exists in various
inter-convertible technical representations, e.g. for communication (transport)
between systems and system components (applications, services, infrastructure
components, process models) or for persistence. A BOs technical representation is
also employed for the passive control of processes and involved systems: decisions
are based on the BOs state and attribute values and the BO basically does not
actively influence its environment.
It is important to distinguish between BO types and BO instances. While a BO
type serves as a blueprint used when designing processes or software (e.g. cus-
tomer), a BO instance denotes an actual, living object build according to a BO
blueprint (e.g. Acme Ltd.).

2.2 Design of the Business Objects Model

To close the horizontal gaps only requires a common canonical model on all layers.
To close the business-IT gap, the BO model is designed to be used by logistics
domain experts as well as by IT-experts and system integrators.
However, since these groups have different focus, they are supported by different
views of the BO model. To narrow the business-IT gap, these views are tightly
coupled and automatic translations are supported, as depicted by Fig. 2.

2.2.1 Domain ModelFunctional View

As described above, the domain model is used by logistics experts. In order to be

easy readable and accessible and also transferable into different representations (e.g.
Java), it is represented as UML class diagrams.
The Domain Model describes different objects (including their attributes) that
occur within the logistics business respectively that are communicated within
logistic processes. This includes information about physical objects (e.g. articles or
load support) and documents (e.g. orders).
Seamless Interoperability in Logistics: Narrowing 83

Fig. 2 Business object model

and Views

BOs are composed of attributes and components, where many components are
standardized, e.g. by CCTS [13] and other UN/CEFACT [14] standards. This holds
for OAGIS [15] (Open Application Group Integration Specication) as well as our
Business Object model. Components themselves are also composed of attributes
and components, and from so-called reference components other BOs can be
linked. Therefore, components are the standardized building blocks of BOs. This is
displayed by Fig. 3.
This also holds for the Java representation and the database representation of
BOs, but here references between BOs are represented naturally via BO-typed
attributes or foreign keys respectively.
The bandwidth of the BO attributes is ranging from features like a unique
number to notes and descriptions, physicals dimensions, various statuses etc.
Currently, the domain model consists of 31 business object types, covering the
following domains:
Order Management: from purchase order to invoice
Facility Logistics: from receiving to shipping
Supply Chain and Transportation: inventory, capacity and route information.

Fig. 3 Inner structure of

business objects
84 M. Bhmer et al.

The domain model consists of more BOs than found in the UML model: The
domain model adds BOs, which only differ in state or verb/method usage, but have
the same form. An example is BO Shipment, which is also used for Notication-
OfDispatch (Shipment [announced]), Fulllment Conrmation (Shipment [dis-
patched]) and DeliveryNote (Shipment [delivered]).
The BO model is well documented in [16], however currently available in
German only. The documentation provides motivations and detailed UML dia-
grams for each BO as well as comments for each attribute in German and in English
language. In the following the BOs PurchaseOrder and HandlingUnit are explained
in detail.
The purpose of the PurchaseOrder (PO) business object is to communicate an
order to purchase goods from a buyer to a supplier (see Fig. 4). The PO is an
evolution of EDI 850 Purchase Order and OAGIS PurchaseOrder. It carries
information to and from the buyer and supplier, including terms of payment and
terms of delivery. The PO is a legally binding document once both parties agree to
the contents and the specied terms and conditions of the order.
A PurchaseOrderLine and a SalesOrderLine are linked with a Warehouse-
ShippingOrderLine. The SalesOrder (SO) is a step beyond a PO in that the receiving

Fig. 4 Business object purchase order

Seamless Interoperability in Logistics: Narrowing 85

Fig. 5 Business object handling unit

entity of the order also communicates sales information. The WarehouseShipping-

Order (WSO) is used to instruct warehouses to ship orders. It is commonly used to
authorize a warehouse to make, conrm, modify, or cancel a shipment or a previous
shipping order. By associating order lines, partial deliveries and combining of orders
in the transition from one order to another one can be realized.
The understanding of the BO HandlingUnit (HU) follows the denitions of [17]
and [18] (see Fig. 5). A HU is a physical unit that consists of load support and goods
contained therein or thereon. HUs are the basic units of all material handling pro-
cesses (put to stock, retrieve from stock, order picking etc.) and associated infor-
mational processes (scan/identify, inventory control, etc.). All the information
contained in the items, such as batches and serial numbers, location where the items
are stored, physical dimensions etc. are available by a HU. HUs can be nested,
meaning that one HU can be created from other HUs (e.g. HU Container ! HU Pallet
! HU Case). They have unique identication numbers that can be formed according
to standards such as GS1-128 or SSCC (Serial Shipping Container Code) [19].

2.2.2 Technical Model Representations

Each BO technically comes in different representations, which technically enable or

focus on different aspects and which are designed to support different architecture
layers. Each of them is machine-interpretable and derived from the logistics domain
model (UML), since we generate Java code from the UML model, and Java is
central for the generation of all technical representations.
86 M. Bhmer et al.

Structure. The structure is dened by the UML domain model, which names the
BOs, their attributes and relations. It denes the structure of the behaviour repre-
sentation and of the persistence representation, while the transport representation is
based on accepted technical standards.
Transport. The main and canonical standard transport representation is OAGIS
[15], namely our light OAGIS overlay and its business object documents (BODs).
However, by using converters, any other market-relevant BO transport format can
be used if a legacy application needs it, e.g. UN/EDIFACT [20].
The transport format is transactional: It does not only consist of data (one or
more BOs), but additionally describes the receiver and what the receiver is
requested to do with the data. OAGIS therefore provides BODs (business object
documents) being exchanged between partners. The BOD structure is displayed in
Fig. 6: Each BOD mainly consists of a verb (what to do with the BO) and one or
more nouns (or BOs, as we call them). The nouns/BOs themselves are composed of
components and elds (attributes). Besides this data area, BODs have an applica-
tion area containing some general information like sender and receiver.
The verbs express what the receiver is expected to do with the nouns. The verbs
come in pairs (request/response) and are also standardized by OAGi. They can in
principle be combined with all nouns/BOs, which reduces the learning curve for
BODs. Get and Change represent CRUD methods; Process is used for business logic.
Notify means using BOs for events and asynchronous communication. Sync is used
for master data synchronization. Figure 7 displays this orthogonal construction.
BOs in transfer format separate state and structure from their behaviour, because
they are communicated between different systems (and possibly also between dif-
ferent layers and components of one system). This holds because behaviour is not
transferable nor easily portable and has to be coordinated flexibly by higher-level
business process models. Emergence (interference of different behaviour descrip-
tions) should be avoided.

Fig. 6 BOD structure (from

[21])
Seamless Interoperability in Logistics: Narrowing 87

Fig. 7 Verbs are paired and

freely combinable with nouns/
BOs

Behaviour. Business Logic is supported by the Java representation of the BO

model. A Java entity model is generated from the UML model. It provides CRUD
methods and more. Thus behaviour information is type information, not transported
with individual object instances.
Persistence. The persistent representation of the BO model is generated from the
Java representation using JPA (Java Persistence API). In this way, different rela-
tional database system and even NoSQL databases can be used to store and
manipulate BO instances. We even provide a BO Instance Repository (see
Sect. 4.1).
A BO therefore exists in different inter-convertible technical representation
formats: the standard transfer format may be parsed into equivalent technical
objects (with methods in their types) and may be persisted in a database if needed.
The persistent representation form is the master copy of a BO and is hosted in a
unique application or service (the objects owner), which is responsible for the
BOs life cycle. All other types of representation (and even other persistent rep-
resentations) are object copies related to the object owner.
OAGIS is mainly used for standardized communication between different tools.
Our BO model, however, and especially its implementation, has a much broader
scope: as stated above, it is also used for Java communication and for persistence.
This implies many difculties in implementing the model, e.g. automatic mar-
shaling and unmarshalling between Java and OAGIS-XML, as well as consistent
strategies for lazy fetching and cascading of create, update and delete operations on
such complex data structures. The BO Instance Repository implements all this
consistently.
Since the OAGIS overlay schema is very complex and uses all XSD features, we
automatically derive a simplied XML schema from the standard schema. This
simpler schema can be used alternatively to construct and validate XML instances
of BOs, without any restrictions. However, it only uses a simple subset of XSD, e.g.
no inheritance, substitution groups, choice and others, and therefore even rather
simple tools can use it.
88 M. Bhmer et al.

2.3 Domain Model Development Methodology

Developing a data model containing all business objects needed in the considered
domain and identifying the existing connections is a very complex task. In
accordance to the basic understanding in systems and software engineering, the
following phase-based and iterative approach proved to be suitable and was applied
for BO4L.
Stage 1: Problem and Requirements Denition. In the course of this stage,
typical workshops with experts of the considered domain should be held. Supported
by the use of creative methods such as concept card and brainstorming, the relevant
requirements should be identied. For details concerning requirements of the BO4L
domain model, e.g. see [12].
Stage 2: Analysis. The goal of this stage is to identify every needed BO within
the considered domain. As predominant standards developing organizations like
Open Application Group (OAGi) or GS1 have shown with their business document
standards OAGIS [15] and GS1 XML [22], the best way to identify business objects
is based on scenarios. However, experience within the team of problem analysts has
shown that, in order to facilitate BO identication, scenarios should be described on
a much more detailed level than e.g. OAGi and GS1 use within their scenarios. This
stems from the fact that not only business-to-business (B2B) communication should
be facilitated, but also communication between business process applications with
one application being just a small fraction of todays range of functions of business
applications, i.e. receiving or inventory control.
Stage 3: Design. Within this stage, identied BOs are designed. In this context,
design means rstly the modelling of the business objects associations and sec-
ondly the denition of all its attributes. Revealing associations is thematically very
closely linked to Stage 2 and may lead to frequent returns and reassessments of its
ndings. Therefore, it seems to be more efcient in terms of design efforts to rst
create associations and then to design the object itself in more detail.
Certainly, the objects attributes are closely related to the scenarios. Again,
scenarios play a key role in this stage. However, even though they are inevitable,
they should be enhanced with information provided by existing standards in order
to gain completeness of attributesas far as possible. In case of the BO4L
development, well-established business document standards (OAGIS [15]; GS1
XML [22]; EDIFACT [20]; openTRANS [23]) were chosen as references. The
working group constructed a model that can be regarded as an evolution of these
standards. In this way, it was assured that the resulting business-IT gap is minimal.
Finally a rst UML model results.
Stage 4: Maintenance. It is clear that no model is comprehensive and faultless
right from the startno matter how smart and thoroughly the design process is.
Maintenance is therefore a key element of every models life cycle. The way this is
managed by model governance will be explained within Sect. 2.5.
Seamless Interoperability in Logistics: Narrowing 89

2.4 Technical Model Development Methodology

The development of the technical model is also a multi-stage process, consisting of

the following steps:
Stage 1: Selecting a Basic Standard. For yielding acceptance of the Logistics
BO model, it has to be based on accepted standards. Several transaction standards
were evaluated, as described in Sect. 3.1, and OAGIS was selected.
Stage 2: Producing a Light OAGIS Overlay. The outcome of the functional
analysis is materialized as an XML schema. The base standard OAGIS is adapted in
a standard-conform way, so that it fulls all our requirements. An extended subset
of OAGIS results, where also the extension consists of standard components. This
construction process has the following steps, as depicted by Fig. 8.
Removing nouns: All OAGIS nouns are deleted, which are currently not needed,
i.e., which do not correspond to a supported BO. These BOs may be reactivated
later, as needed.
Construction of the Overlay by standard means: attributes needed, but missing
in OAGIS are added mainly by sub-classing the OAGIS noun. There is an OAGi
guideline document on how to do this, see [24] and its addendum [25].
Moreover, even new nouns with all their needed attributes are added as needed.
Sample instances: for all components and all BO small sample instances are
constructed bottom-up, containing only the supported attributes. This process is
guided by a structural containment graph, constructed beforehand using algo-
rithm developed earlier. The samples need not contain meaningful data, but with
meaningful data they constitute an important artefact for users of the technical
BO model.
Lightening the Overlay according to the small sample instances: by using the free
tool XMLHelpLine, the XSDs of step 2 are reduced to contain only those de-
nitions available in the sample instances from step 3. This is a fully automatic
process. Optionally, the tool can also do a flattening, reducing the number of les
resulting.

Fig. 8 Construction of the technical business object model as a light OAGIS overlay
90 M. Bhmer et al.

Stage 3: Constructing a better Domain Model. After these steps we had two
model representations: the initial UML model from the functional analysis and the
OAGIS overlay. Our goal was to use the UML model also for generating other
technical artefacts, namely the Java entity model. In Sect. 3.1 we describe that the
construction of the Java entity model is a manual engineering step. It could not be
derived from the technical OAGIS representation, but the UML model has been
redesigned to get all model representations in harmony as much as possible, and to
use automated steps for the development.
Stage 4: Artefact Generation. Many of the BO model artefacts are constructed
automatically: The technical model is based on an automatically lightened OAGIS
overlay, complemented by a Java entity model generated from the domain UML
model, which is therefore a business-oriented Java entity model. From the Java
model, a database model is generated using JPA, used by the BO Instance
Repository. Moreover, BOD XSD les and WSDL les are generated based on
templates and conguration les.

2.5 Model Governance

The set of BOs and their associations form a consistent data model of the logistics
domain. As stated above, it is impossible to represent this complex industry with all
its sub-disciplines completely. These circumstances led to a scenario-based
approach for dening BOs (see Sect. 2.3).
Furthermore, practical experience proved that such domain models are subject to
change as frequently as the domain itself. Hence, a fundamental assumption is that
the BO model is experiencing permanent development. Dealing with the change
becomes a crucial challenge.
Model governance is mainly driven by the functional model. Besides the
functional model, several other aspects are affected by changes, i.e. documentary
reports and technical models. OAGi describes its development methodology in
[26]. For the technical XSD part, this has largely been adopted. The developed
process model for the evolution of the functional model as driving force for further
development is explained in the following.

2.5.1 The Ecosystem

The model governance is incorporated into a technical ecosystem, dening the

concept of governance, identifying its inner actors and technical components and
linking those by activities. Several policies apply to this ecosystem like a versioning
specication or process instructions for changing the denition of a BO Fig. 9.
Resources can be subdivided into human actors, the BO model, and (IT) com-
ponents. Because all roles listed in Fig. 10 work with those BOs, they are incor-
porated into the ecosystem as users (stakeholders). Two more groups of actors have
Seamless Interoperability in Logistics: Narrowing 91

activities
BO
Obj3
-attr1
technical
-attr2
view

BO BO
Obj1 Obj2
functional
-attr1 -attr1
view
-attr2 -attr2
(IT) components model actors

n
io
p o l i c i e s

ut
ol
ev
Business Objects Ecosystem

Fig. 9 Business Objects Ecosystem

Fig. 10 Governance process of the domain model

been identied: people who actively develop the model (developers) and a board
deciding in case of doubt or conflicts. Without such a board, the process of
development would certainly become ineffective due to the heterogeneous partic-
ipants and their conflicting intentions. The BO model affects several (IT) compo-
nents, which in most cases are artefacts of the overall Logistics Mall. The functional
and technical views on the model are of great signicance in this context because of
their mutual dependencies.
Interactions among resources are represented by combining specic resources
with an activity to form a use case.
92 M. Bhmer et al.

2.5.2 Evolution of the Domain Model

Everything in the ecosystem will change over time. These changes occur on two
different levels: Modications of the BO model or their dependent components are
considered substantial (content level). Controlling these is the main purpose of the
ecosystem. However, in case a policy changes or when a new activity is added,
change takes place on a meta level.
The ecosystem can control itself by corresponding (meta) policies, e.g., a policy
on modifying content-level policies. Because the focus certainly is on the evolution
of the BO model, the developed change process is presented here (see Fig. 10).
This BO model maintenance process has the following steps:
Phase 1: Change Request. The process already starts with the necessity to
change the model. This can be expressed by every arbitrary combination of the
following actions involving one or more BOs: Adding a new attribute/association,
modifying an existing attribute/association, deleting an existing attribute/associa-
tion. All developers within the ecosystem (see above) are eligible for submitting
their suggestions to the board.
Phase 2: Assessment of Requests. The Governance Board checks the technical
and functional compatibility with the existing BO model. Besides compatibility, the
decision on the approval can be influenced by strategic and economic aspects.
Changes could be denied because of functional or technical weaknesses, strategic
aspects, or economic aspect, e.g. shortage of implementation resources. This could
then lead to adjournment and resubmission. In both cases (denial and adjournment)
the requesters are informed.
Phase 3: Realization and Quality Assurance. In case of approval, the changes
are implemented into the BO model. The process of implementation is again subject
to policies, containing rules and maxims, set up for a consistent and high-quality
result, independent of the individual who performs the changes. The main result of
this phase is a new BO model release. Due to the dependencies, new (versions of
the) components are produced as well. During this phase, the board checks the
compliance of the implementation regarding the policies of phase 2. In the case
when discrepancies are found, phase 2 is performed again. When the new model
and the components are nally approved, they can be published.
Phase 4: Publication. A new release of the BO model is made available to the
users of the ecosystem. However, not every implemented and approved change
request will result in a new release. Changes performed within a certain period are
combined into a single release. There are a couple of important aspects to be
considered for the process of evolution not mentioned here. All of these are covered
by dedicated policies. Examples are versioning and the lifecycle of a release or
version. They are introduced in the subsequent section.
Seamless Interoperability in Logistics: Narrowing 93

Fig. 11 Levels of versioning

2.5.3 Versioning Concept

Versioning is a big issue within the topic of Model Governance. On the one hand,
versioning should facilitate evolution. On the other hand, it must guarantee stability and
reliability, as these are crucial for its application and success in practice. Therefore, any
model governance should come with an adequate concept of model versioning.
Especially, the aspects of backwards and forwards compatibility need detailed atten-
tion. Other aspects are version lifetime and the number of parallel valid versions, since
these effects in a large part costs for maintaining of several old versions.
Figure 11 shows three levels of versioning concepts considering these con-
straints. Revisions on the lowest level (patch level) and the medium level (minor
release) are backwards compatible. While changes on patch level only include
smaller optimizations and bug xing, minor releases include adding of new BO and
BO attributes and technical changes. Major releases allow substantial changes of
the domain model, e.g. changes in structure and deleting attributes and objects.
As major releases are not backwards compatible, migration efforts on App-level
have to be carried out, i.e., all apps using these BOs in their new mayor version
have to be modied. Thus, major releases are produced once a year and a lifespan
of ve years seem to be appropriate (see Fig. 12).

Fig. 12 Lifespan of domain model versions

94 M. Bhmer et al.

3 Narrowing the Business-IT Gap Using Business Objects

This section describes how the business-IT gap and the two horizontal gaps men-
tioned in the introduction are closed concretely, i.e. how the vision initially sket-
ched by Fig. 1 is realized.
For lowering the horizontal gaps a common model is used, this is the domain BO
model for the business-business gap and its canonical1 technical representations for
the IT-IT gaps. Thus, the BO model is a common language with two dimensions:
business and IT. The business dialect is understood by business people and the IT
dialects are understood by different layers of IT systems.
In Sect. 3.1 we discuss the horizontal IT aspect, namely canonical application
communication, orchestrated by protocols. The horizontal domain aspect was
already discussed in Sect. 2.
In Sect. 3.2 the business-IT gap is closed on data level by mapping the business
view of BOs to its IT view. Then Sect. 3.3 closes the business-IT gap on activity
level, where domain tasks are mapped to IT system APIs if they can be automated-
manual tasks remain. Section 3.4 nally connects all aspect on the business process
level and summarizes the argumentation.

3.1 Horizontal IT Gap: Canonical Application

Communication

Logistics services or applications like a warehouse management system, a transport

management system or an ERP (enterprise resource planning) systems and even
smaller logistics modules and services, which we call apps, are all seldom used
stand-alonethey have to cooperate, e.g. by exchanging messages. Interoperation
is needed for the different applications of a customer all available via the Logistics
Mall (i.e., in the Cloud) and for the apps of his cooperation partners in the Mall
Cloud, but also between applications in the Cloud and applications the customer
already uses and still wants to use on premise. This requires a common data model
to reduce integration efforts.
The Logistics BO model is canonical [27] and independent of any software
product or software provider. Even more, it is based on accepted technical stan-
dards constructed in decades by technical (and domain) experts from the standards
organizations and their supporters, and it has an abstract domain view constructed
and discussed by logistics domain experts from two Logistics project clusters.
Therefore, it supports the integration of different applications in an optimal way:
For every application, only one business model adapter has to be developed,
independent of the number of its communication partners, as shown in Fig. 13 on

1
canonical means conforming to accepted principles or standard practice, i.e. being harmonized,
agreed on, minimal, standardized, reusable, but also extensible [27].
Seamless Interoperability in Logistics: Narrowing 95

Fig. 13 Canonical communication via standard business objects

the right side. This contrasts with the left side, where for each communication
partner a different interface has to be provided and maintained, and different lan-
guages are used and many mappings are needed, since same things are named and
structured differently.

3.1.1 Selecting a Base Standard

Widely accepted e-business standards are central for interoperation of applications,

and they are used on top of technical interoperation standards. Not only the business
objects exchanged need to be standardized to ease communication with different
kinds and instances of other applications, the protocol steps have to be considered
as well.
Today, many e-business standards exist, as described in Sect. 5. We compared
and evaluated different transaction standards and selected OAGIS [15], the Open
Application Group Integration Specication. The reasons are:
it is a powerful, adaptable and extensible horizontal standard, based on Core
Component Technical Specication CCTS [13] and many other standards
it is mature (rst version in 1996pre-XML, we used version 9.5.1) and in real
use (broad industrial support in many countries)
a logistics module and many logistics-related denitions already exist (e.g.
container, shipment, item, pick list), only few extensions are needed for logistics
it is technology-neutral, but includes Web service denitions, and there are
OAGI working groups on Cloud, Mobile and BPMN support.
UN/EDIFACT [20] formats (in widespread use) are convertible to OAGIS.

3.1.2 Benets of Using Canonical Standardized Communication

The benet of using such a widely accepted standard for canonical communication
is that the format of most messages ever needed is already dened based on the long
experience of a large expert group. Only few denitions have to be added, based on
96 M. Bhmer et al.

Fig. 14 Cost benets of

canonical communication

the standards extension mechanism. These denitions are in turn based on CCTS
[13] and its Core Component Library CCL, a mature reservoir of components used
as semantically described building blocks in OAGIS and several other standards.
Thanks to its many optional attributes, even simple messages can be exchanged
conformant to the standard. If special information is needed, its position and format
is already dened, thus minimizing data mapping needs.
This enables stable, standards-based interoperation interfaces of applications or
services, initially implemented as needed (lling only needed data elements) and
subsequently completed as needed in a downwards-compatible way. The same
standardized interface can be used for many different partners, and additionally EDI
converters can be provided: as OAGIS itself is based on UN/CEFACT standards,
transformation to its grandpa UN/EDIFACT still widely in use in logistics is
possible and supported.
As depicted by Fig. 14, when using a canonical model, the cost for integrating a
small number of apps is initially higher, but only increases linearly with the number
of interoperating apps. Especially when modications of apps occur over time, the
canonical communication is benecial.

3.1.3 Technical Communication via BODs

OAGIS denes several BODs (business object documents) exchanged between

applications. Each BOD is composed of an Application Area (the document header)
and a DataArea, consisting of a verb and one or more nouns. The verb denes the
action to be performed on the nouns or data, which represent the main BOs, visible
in the application or service interface.
For each noun, there is a service having methods representing meaningful verbs
of that noun. This service can be implemented by various technologies, e.g. RMI
(Java Remote Method Invocation), REST or as a Web Service. BODs are used as
message formats for these methods, regardless of the technology used. OAGIS
itself comes with sample WSDL denitions (the OAGIS services), but we added
and implemented services for RMI and REST.
Each application or service running in or outside of the Logistics Mall requires a
Mall interface consisting of its required OAGIS services in a technology. This
interface has to be implemented towards the enterprise service bus (ESB) used in
Seamless Interoperability in Logistics: Narrowing 97

the Mall or towards the BO Repository. It can then communicate with other
application also implementing such standard interfaces. Besides this standard
communication, peer-to-peer-communication using other formats is still allowed to
smooth the integration curve, e.g. by using EDI converters.
For efciency reasons, the Logistics Mall internally often uses call by reference.
For this purpose, we dened MiniBODs, which are normal BOD, but mainly
contain identiers of BOs instead of copies of BOs. These MiniBODs uniquely
reference BOs stored in the BO Repository.

3.1.4 Communication Protocols

To orchestrate the application or app communication, different scenarios and pro-

tocols can be realized based on the asynchronous exchange of BODs, as depicted
by Fig. 15, an UML sequence diagram. Here, some customer orders goods from a
supplier, and for the shipment, a logistics carrier is used. The process starts by

Fig. 15 OAGIS logistics scenario no 59-Extract of [15]

98 M. Bhmer et al.

sending a ProcessPurchaseOrder BOD message from customer IT to supplier IT,

and the latter acknowledges the order using a BOD. Communication proceeds as
needed, until nally the invoice is sent to the customer IT and acknowledged. Such
diagrams will soon be available in BPMN from OAGi.
As stated above, these BODs are message formats in OAGIS services, which
constitute the OAGIS application interface services of involved parties. Such pro-
tocols are often hard-coded in the applications, where receiving a BOD will even-
tually result in sending the answer BOD, either synchronously or asynchronously.
A more flexible and more abstract orchestration of services can be dened using
business process models, as described in Sects. 3.3 and 3.4. As a preparation, we need
a mapping of the domain business objects model to its technical representations.

3.2 Vertical Data Gap: Mapping the Domain Model

to the Technical Model

As OAGIS is our underlying base standard and OAGIS is dened by a set of XML
schema les (XSDs), we rstly evaluated several methods to construct an UML
model from these XSD les automatically. However, it turns out that this is not
feasible to construct a domain modela good abstraction is a manual engineering
step, and the mapping has to be stored somehow.

3.2.1 Automatic Mapping Approaches

OAGi itself used the free Eclipse plugin HyperModel for creating UML diagrams
for their denitions. HyperModel is able to import XSD les and to display them as
UML diagrams, and it supports some operations on the diagrams, e.g. merging
inheritance. However, the resulting UML is very complex and not at all under-
standable by domain experts. The UML diagram for a single business object would
ll a wall if expanded in all detail. See Fig. 16 for a sample, namely a Purchase-
Order (still with unexpanded parts). This even holds for our reduced overlay, since
HyperModel displays all the details in any depth.
The next approach was to generate Java from XSD, and then to generate UML
from Java. There are several tools for both steps. The rst step is critical and
supported, e.g., by tools like XMLBeans, JiBX, JAXB, CXF WSDL2Java, Castor,
Smooks, AndroMDA and others. The interested reader may easily nd references to
these techniques and tools. Although some tools had difculties with some con-
structs used in the OAGIS XSDs, it worked, but the results were again very
complex-the resulting Java model was rather unusable. Also for the next step, the
mapping of Java entities to database schemas to implement BO persistence, dif-
ferent OR-mapping techniques were investigated (DataNucleus, JDO, JPA, etc.)
and even NoSQL approaches, but all of these techniques require a Java model
simpler than generated.
Seamless Interoperability in Logistics: Narrowing 99

Fig. 16 Inappropriate domain model: Part of the OAGIS purchaseOrder noun

3.2.2 Engineering the Mapping

Therefore, we understood that constructing a good domain model is a manual

sophisticated engineering task, requiring excellent domain knowledge, even when
being directed by a highly detailed and elaborated technical model like OAGIS.
Sophisticated abstraction is needed to hide the complexity (which is needed for tool
integration) from the domain expert, and meaningful abstraction cannot be gener-
ated. On the other hand, the abstractions constructed should be compatible to the
original model, since a simple mapping from the domain model to the technical
model is essential for our approach, as described in Sect. 3.
Figure 17 illustrates the construction of the logistics domain model as an
abstraction of the subset used from the technical model: unused attributes are
ignored and long paths are compressed. New nodes are sometimes inserted to
regroup the result and increase comprehensibility. The abstraction was developed
by several domain experts; therefore, it is understandable for logistics domain users
having no technical knowledge. The result for the BO PurchaseOrder is shown by
Fig. 4 in Sect. 2.2. This simplied model has several good properties:
It is simple enough to be understood by domain users
It is simple enough for further generation steps, e.g. for the generation of per-
sistence support
It can be extended by methods, i.e. adding behaviour to the structure, at least
CRUD methods (Create, Read, Update, and Delete) with some sophisticated
aspects like CRUD on collections and managing inverse relationships.

3.2.3 Maintaining the Mapping

Each attribute in the UML model is mapped to the complex standard OAGIS XML
path, stored in the BO Metadata Repository (see Sect. 4.2). The resulting benets of
this construction are as follows:
100 M. Bhmer et al.

Fig. 17 Mapping the domain model to the technical model

The domain expert has a simple and understandable view of an underlying

complex model and can use it e.g. in process modelling.
The IT expert has a detailed technical model and even several technical repre-
sentations, capable for standard-based tool integration and communication with
other tools.
Both models are tightly coupled and automatic translation is possible
The business-IT gap is minimized, since the domain model BOs are pre-
implemented in Java and as standard OAGIS-based XML schemas. The Java
mapping is 1:1 UML to Java generation, while the OAGIS mapping is 1:1 stored
paths. Domain artefacts are pre-linked with different technical artefacts, as
described in Sect. 3.3.

3.3 Vertical Behaviour Gap: Pre-implemented Activities

for Process Models

The business-IT gap not only exists for data, but also for handling and manipulating
data, i.e., for business tasks performed by IT systems, if the tasks can be automated.
The business tasks deal with the business view of data, while the IT systems need
technical representations. Process models can be used to orchestrate business tasks,
not only automatable tasks but also manual tasks, executed with the help of soft-
ware or purely manually with and on logistics BOs.
Seamless Interoperability in Logistics: Narrowing 101

3.3.1 Business Process Models

In times of international competition and tight prot margins, efcient business

process management including partner interactions is essential for logistics com-
panies, especially for SMEs. Since prices are dictated by the market, only the
internal optimization of process models can help setting some unique selling points:
Only individual process models, not the standard ones, can make a difference in the
market: the unique selling points.
Business demands change quickly, especially in logistics. The business pro-
cesses of a company should be quickly adaptable to new requirements. However,
often a quick adaptation is not feasible, as the implementation of new logistics
process models, designed by logistics experts, also requires additional IT support
that takes time to be brought in place.
To speed-up process design and implementation, the Logistics Process Designer
LPD [28], as part of the Logistics Mall, enables quick process adaptations without
the need for in-depth IT expertise and enables the deployment of these process
models merely by a click of a button. It is based on standard BPMN 2.0 [29],
generated from a simpler domain-oriented process model, and used for process
execution in the Cloud as well: the business users can focus on their business
processesthey do not need to care about the process engine and IT concerns.
Business processes are modelled by domain users, not by IT experts; therefore,
process models focus on control flow. However, data flow can also be modelled: in
domain view, activities have inputs and outputs, both described by some BOs from
the domain model. The inputs of an activity have to be produced by some prede-
cessor activity.

3.3.2 Technical View on Processes

Technically, each process instance (being a started process model) has its local
process context, and each activity in the process instance can access and write data
in its process context. Activities are implemented by API methods of IT systems or
by special services or apps.
To minimize the business-IT gap in process modelling and process execution
means to minimize process implementation efforts, since normally process imple-
mentation needs IT experts. Here we use building blocks on domain level as
depicted by Fig. 18, which are pre-implemented by IT system methods and used as
a modelling unit.
These building blocks have to specify how the business activity communicates
with its IT implementation. This is eased by pre-implemented BOs used in the
domain view and pre-mapped to the technical view. They close the business-IT gap
for data and for behaviour.
Such a block mainly consists of additional metadata for IT systems and apps
offered via the Logistics Mall. The metadata describes the names, inputs, outputs
102 M. Bhmer et al.

Fig. 18 Pre-Implemented activities as building blocks on domain level

and behaviour of all activities which can be performed using the offered system and
is used to generate BPMN 2.0 XML task descriptions.
Since our Logistics BOs are rather complex and may become very big, we
relieve the process instances (or better: the process engines process contexts) from
handling themthe process engine only stores IDs (i.e. references) of BO instances
which are stored in the BO Instance Repository of the process owning party. In this
way, the IDs exchanged between process instances and apps coordinate their data
transfer with the BO Instance Repository, while the processes organize the control
flow between apps.

3.4 Narrowed Gaps: Paradigm Shifting in Process Modelling

The canonical BO model is central for the orchestration of Logistics Mall products
(apps) by using business process models, designed and deployed by business
people (with minimal IT support). The mechanism allows phase inversion, since it
is based on pre-implementations of two kinds of modules:
Behaviour modules: Activities for offered Logistics Mall products using BO-
based interfaces, as described in Sect. 3.3
Data modules: standardized product-independent BOs, as described in Sects. 3.2
and 2.2

3.4.1 Phase Inversion

Most BPM suites have different tools for process modelling by domain experts
(phase 1 in Fig. 19, modelling a business process using activities A1, A2, etc.) and
for the later implementation of these process models by IT personal (phase 2,
implementing each Ai by some implementation Ii, e.g. by a service call). However,
in the Logistics Mall we deploy and execute modelled business processes merely by
a click of a button, not involving IT knowledge or IT personal.
Seamless Interoperability in Logistics: Narrowing 103

Fig. 19 Phase inversion-conventional approach versus logistics mall approach

This is only possible, if we model with already implemented building blocks

with standard interfaces using standardized data formats, since implementation after
modelling should be minimized. In this case, we have a phase inversion, also
depicted by Fig. 19: the application or service provider (i.e. the provider of I1, I2,
etc.) not only provides normal metadata, but also process model-related metadata
for his offer (phase 1). This serves as implementations of activities used in process
models, since from this metadata, activity descriptions in BPMN 2.0 XML can be
generated. Process modelling and modications of process models are then the next
steps. They have no further influence on activity implementation. The same activity
implementation can be used in different process models of the same or of different
tenants, and in any new versions of these process models as well. The modelled
process can be changed and is always directly executable, since it uses pre-
implemented building blocks using standardized pre-implemented data models in
its interfaces.
This principle is sometimes called What You Can Model Is What You Exe-
cute in analogy to the WYSIWYG principle of text editors. Here the word Can is
important, since of course this modelling approach has some restrictions, in analogy
to Lego construction sets.
BPMN 2.0 optimally supports this idea, since it uses the same notation for
process modelling and for process model deployment and execution. The latter has
additional attributes set in its standardized exchange format BPMN 2.0 XML. Here,
the task description with all implementation-relevant attributes of the process-rel-
evant service metadata is inserted for the modelled tasks.
Moreover, the idea is also practical in the context of the Logistics Mall, since the
scope of its process models is limited to the logistics domain, limited in complexity
(if complex logic is seen in the applications) and even stronger limited to only
orchestrate concrete applications and services offered via the Logistics Mall or
made accessible to the Logistics Mall. Basic services (such as a mail service) will
be offered free, to get a critical mass of building blocks. In [30] it is described how
top-down modelling can be achieved.
104 M. Bhmer et al.

Nicholas Carr [31] wrote IT Doesnt Matter: companies cannot achieve

competitive advantage only by IT, since IT is no limited resource and other com-
panies may use similar IT (but perhaps not SMEs due to the cost factor of IT). Such
benets only arise because a company does something better than others, con-
cerning their business strategies. The approach presented here refutes this argument,
since companies may map their respective business peculiarities of various
departments directly to IT flexibly and quickly. Any IT, which flexibly allows
realizing and changing core business strategies, does matter.

3.4.2 BO Support for Phase Inversion

The logistics business object model BO4L follows the same principle in its two
views. There is an abstract domain view dened by UML diagrams, which hides the
complex details of the underlying OAGIS Logistics Overlay. The abstract model
was constructed based on the technical standard (again some kind of phase
inversion) by several logistics experts of Fraunhofer IML and in conjunction with
other Leading-Edge Cluster logistics projects, to assert both the renement relation
of the two models and at the same time the understandability by domain experts.
While the technical model is in English only, both the domain model and the
application or service metadata and process metadata are multilingual and com-
mented (descriptions can be displayed in the process modeller in a selected lan-
guage). The domain views are only used during process modelling or service
provision, while the technical models are used at runtime.
We call this the double barbell principle [30], since for both, behaviour (the
process model) as well as data structure (the business objects), the corresponding
domain view and technical views are closely related in advance like a barbell
(compare Fig. 20). The technical views cannot be constructed as a later renement
(of arbitrary BOs or activities), since IT personal should not be involved. Instead,
the barbells are constructed only once by the providers of applications (and the BO

Fig. 20 Barbells pre-connecting domain view and technical view of data and activity
Seamless Interoperability in Logistics: Narrowing 105

model governance team) and then reused in any process models. The standard BO
model is important to avoid complex data mappings, as the standard activity model
is to avoid complex implementation. Both are done only once, not for any new
process model or model changes. This is especially important in a Cloud envi-
ronment, since Web interfaces for the implementation tasks are difcult to develop.
Compared to Fig. 1 in Sect. 2.1 the GUIs of IT applications or apps are replaced
by activities, which may be manual activities implemented by the GUIs or auto-
matic activities implemented by the IT systems or apps.

3.4.3 Putting Everything Together

Figure 21 depicts the resulting architecture with all gap reductions: The upper part
is business: process models are used to describe business processes handling
business data, indicated by dashed lines in process models. Data flow in process
models can be specied, but in principle is bus-likeevery BO produced previ-
ously in the running process can be accessed (from the process context). The
process models together with a common data model, the Logistics BO model,
bridge possible business-business gaps, e.g. between cooperating partners.
The bottom part is IT: different apps (i.e., implementations of activities) com-
municate via technical BOs in Java or OAGIS XML representation, and here we
use the BO Instance Repository as a counterpart to the process contexts of all
process instances maintained by the process engine. The process instances only

Fig. 21 Narrowing vertical and horizontal gaps

106 M. Bhmer et al.

handle BO-IDs, which are used by the orchestrated apps to access and exchange
data. This closes the IT-IT-gap.
The most important part is closing the vertical business-IT gaps for data as well
as for each activity. This is done by pre-implemented BOs and pre-implemented
activities generated from the Logistics Mall offers and operating on BOs. These
connected building blocks composed of a business item and an IT item are flexibly
orchestrated by business process models, resulting in a What You can Model is
What you Execute approach.

4 Software Components for Seamless Integration

The business-IT gap is narrowed even more by proving supporting components that
ease BO model handling and seamless integration. It is important for the approach
presented in the previous sections to get support for BOs during design time (e.g.
when designing business processes or when registering Logistics Mall app offers)
as well as during runtime (i.e., when executing business processes and apps). For
this purpose, we offer the BO Metadata Repository and the BO Instance Repository.

4.1 BO Instance Repository

Having a common data model (i.e., the BO model) used by all tenants and all apps
is a central idea of the Logistics Mall. Then, providing a logically common data
store for these BOs is a logically next step. This BO Repository also provides a
generic BO browser.
Preferably, all apps should retrieve data (BOs) from and manipulate data residing
in one central common data store in which all the data is stored only once, to get rid
of the distribution and duplication of data in many data base silos and in many
different formats, which hiders a maturation of IT systems signicantly. C.J. Date,
one of the founders of the ER model, wrote the following in his best-seller An
Introduction to Database Systems [32]: An enterprise should store its operational
data in an integrated database to provide the enterprise with centralized control of
its operational data. At least there has to be a unied consistent view of data.

4.1.1 A Motivating Scenario

Logistics applications like an ERP system, a WMS or a CRM system are seldom
used stand-alone-they have to communicate. i.e., exchange master data as well as
transactional data, not only the systems of one tenant, but also those of his coop-
eration partners, being in or outside of the Mall. This is depicted by Fig. 22: three
Seamless Interoperability in Logistics: Narrowing 107

Fig. 22 Scenario motivating

the BO instance repository

logistics applications are integrated as offers of the Logistics Mall and rented by a
tenant, and they cooperate with an ERP of another tenant.
The application providers make their offers Mallconformantfor existing
systems, they write adapters for standardized BO communication using the Mall
bus system. They have two choices: exchange BOs over the bus, or use the BO
Instance Repository as a common data store, for all own apps or even for apps of
communication partners. The latter choices have additional advantages, especially
for newly developed apps.

4.1.2 Objectives and Requirements

The objectives of the BO Instance Repository are as follows:

Supporting App Integration: The BO Instance Repository can easily be
integrated into existing logistics applications or Apps. Alternatively, the data tier
of apps is given by the Instance Repository, which pre-implements the stan-
dardized BO schema.
Data Consistency: All Apps have access to up-to-date and consistent business
object instances.
Security: All Apps can only access or modify data for which they have cor-
responding rights.
Saving Costs: The integration of apps is eased, data synchronization is auto-
matic, and any unnecessary data conversion is avoided.
Cloud-Based: the Instance Repository must t seamlessly into the architecture
of the Logistics Mall cloud.
Distributed: to avoid possible bottlenecks of a centralized data source, the
components of the repository architecture can be distributed to different nodes.
Model-Completeness: each Instance Repository is able to store instances of all
supported BO types (i.e., it covers the complete BO model). This supports
integrity and consistency.
108 M. Bhmer et al.

Multi-Tenant Support: All Logistics Mall customers have to be cleanly sep-

arated. Here we use private Instance Repositories for each tenant. This eases
migration, security, backup, and scalability.
Multi-Version Support: This has a type aspect and an instance aspect:
BOs from different major versions (releases, see Sect. 2.5) of the BO model
are always stored in separate Instance Repositories, while patches and minor
releases use the same Instance Repository. This eases data migration.
Changing a specic BO instance will result in creating a new version of this
instance if the BO type is versionedotherwise this change is only logged.
Normally, such central data repositories are used for master data only. In contrast
to conventional master data management, we may not only store master data and
structural data in this repository, but also dynamic data and transactional data. The
following benets result: BOs of any kind can be exchanged by reference (not
necessarily but copy), all (access and) changes are automatically logged, and the
implementation efforts are minimized.

4.1.3 Architecture

The BO Instance Repository follows a multi-layered architecture as depicted by

Fig. 23: The central part is the BO Entity Model, which is synchronized with the
BO model in UML (the Java code is generated and can be reverse engineered with
the UML tool).
The UML model even contains annotations to congure the database access
layer (JPA annotations). This layer is independent of the storage technique used:
although relational database systems are a rst choice to store the data physically,
JPA even supports cloud databases and other techniques.
On top of the entity model there is a small business logic layer supporting the
creation, update and deletion of BOs as well as reading and different methods of

Fig. 23 Architecture of the BO instance repository

Seamless Interoperability in Logistics: Narrowing 109

searching BO. This functionality can be made available via different interfaces
(APIs, application programming interfaces), e.g. a Java RMI interface, a REST
interface or SOAP Web services, and additionally a BO browser (GUI), which
allows to inspect every detail of the content in the BO repository and to search for
information. BO manipulation support via the BO browser is planned.

4.2 BO Metadata Repository

To close the business-IT gap even better, we provide a software component that
stores all BO model representations, i.e. the business representation and all tech-
nical representations with detailed metadata in one repositorythe BO Metadata
Repository. It also stores all mappings and it is used by design time tools as well as
runtime components of the Logistics Mall.

4.2.1 Objectives and Requirements

The objectives of the BO Metadata Repository are as follows:

Support for Multilingual BO Model Browsing for Design Tools in the Logistics
Mall, especially for the LPD (Logistics Process Designer [28]) and for App
Development [33]
Search and browse structured metadata for the BO model, including inter-
nationalized names and descriptions of BOs and all their attributes, as well as
of allowed enumeration values and their meaning
Display in simplied view (e.g., flattening of inheritance hierarchies)
Further Support of App Development
Download of implementation les (XSD and Java archives) for the BO
model
Management of mappings between the BO representations (domain model
and technical representations as OAGIS-XSD, Java)
Support of Integration and Communication
Check validity of BO instances
Support of BO Model Administration
BO model versioning
Management/upload of structured metadata for a BO model version
Generation of (domain) documentation for the BO model version
Dependency analyses (which BO used which components and refers to
which other BOs)
110 M. Bhmer et al.

The BO Metadata Repository is able to store metadata of arbitrary UML models,

since its stored metadata can be imported from any UML model of the UML tool
used. This also eases the management of BO model metadata.

4.2.2 Architecture

The architecture of the Metadata Repository in principle follows the same archi-
tectural principles as described for the Instance Repository (cf. Fig. 23). This yields
the same advantages for tool integration.

5 Background and Related Work

5.1 Business Objects

The provision of standardized business objects is not a new topic; therefore, we

give an overview of existing approaches usable in the logistics domain.
It is highly important for a sophisticated BO model to have a good foundation by
using an accepted standard, since Fraunhofer itself, although being neutral, is not a
standards organizationthus we restrict our overview to existing standards. These
standards have been developed over years by several domain experts and many
experiences and corresponding feedback has been considered in these standards.
This fact highly increases acceptance. The main conflict in the selection of a
suitable standard is its widespread use in logistics versus its flexibility and thus the
technical suitability for the Logistics Mall as part of a SOA and for flexible exe-
cuted business processes.
There are three classes of e-business standards to consider:
Process standards (to automate complex business processes) such as ebXML
[34] and RosettaNet [35] with their implementation architectures are quite
complex and mainly suitable for larger enterprises.
Transaction standards (as the basis for the automation of business processes)
like EDIFACT [20] (and EDIFACT subsets such as GS1 EANCOM), GS1
XML [22], UBL [36] and OAGIS [15], but also openTRANS [23] and the
transaction part of RosettaNet appear to be suitable.
Catalogue exchange formats (for electronic delivery of master data) may be
needed additionally, if the process or transaction standard does not support this
task.
Therefore, we focus on transaction standards. There are some comparisons of
these standards from different views in [3741].
UN/EDIFACT (United Nations Electronic Data Interchange For Administration,
Commerce and Transport, ISO 9735) [20] initially developed in the 1980s is an
Seamless Interoperability in Logistics: Narrowing 111

important international multi-domain standard for logistics, since most existing

logistics software provides some EDIFACT exchange interface. It is used by over
300.000 companies worldwide. However, it is technically outdated (delimiter-based)
and thus not future-safe. It is inflexible, nearly not extensible, and targeted to xed
business processes. There are about 223 message formats (BODs) which are ASCII-
encoded (not XML-based), and BODs are not distinguished from BOs. Because of its
complexity, many industry-specic subsets of EDIFACT have been developed
(EANCOM, EDIFICE, Odette, and others), and regularly two versions are produced
each year. Due to its delimiter-based format, it is not well suited for a SOA.
The UN/CEFACT [14] e-Business stack consists of several modular specica-
tions, namely CCTS and CCL: CCTS (Core Component Technical Specication
[13], ISO 15000-5 and part 8 of ebXML) enables cross-organizational interopera-
bility on data level. It is implementation-independent and includes a methodology for
the development of context-specic semantic units and blocks of context-neutral
templates for business vocabulary. An industry-specic adjustment of the data is
performed by a context-driven principle for the generation of variants. Semantic
syntax-agnostic building blocks represent general business data entities (context-free
building blocks, common understanding by machine and human). The CCL (Core
Component Library, TBG17) [42] is a repository for generic, CCTS-based business
information. Core components from the CCL are also called ebXML Core Compo-
nents and used in several eBusiness standards, e.g., UBL, RosettaNet, and OAGIS.
ebXML (Electronic Business using XML) [34] started 1999 is a family of XML-
based standards sponsored by both OASIS and UN/CEFACT and adopted by ISO
whose mission is to provide an open, XML-based infrastructure that enables the
global use of electronic business information in an interoperable, secure, and
consistent manner by all trading partners.
RosettaNet [35] is a non-prot consortium aimed at establishing standard pro-
cesses for the sharing of business information (B2B), working to create and implement
industry-wide, open e-business process standards. These standards form a common e-
business language, aligning processes between supply chain partners on a global
basis. RosettaNet is a subsidiary of GS1 US and has several local user groups. The
European User Group is called EDIFICE. The RosettaNet standard is based on XML
and denes message guidelines, interfaces for business processes, and implementa-
tion frameworks usable for interactions between companies. It focuses on the supply
chain area, but also manufacturing, product and material data and service processes
are in scope. The RosettaNet PIP directory consists of about 30 e-Commerce Message
Types. In Europe, RosettaNet is not widely used, since EDIFACT is mainly used here.
openTRANS [23] from Fraunhofer IAO and the university of Essen is a small
and XML-based transaction standard and can be seen as a complement to the
catalogue interchange format BMEcat [43]. The scope of openTRANS is limited to
eight transaction types: advice, invoice, order, order change, order conrmation,
offer, receipt conrmation, request quote.
GS1 XML [22] from GS1 Germany exists since 2004, it is the successor of
EANCOM, and the most used EDIFACT subset designed for Supply Chain Inte-
gration and therefore has the same focus. It is used by 20 industries in over 20 states
112 M. Bhmer et al.

(2008), but mainly known in Germanyalthough not widely used here, the main
stakeholders include Oracle, Siemens, Phillips, SAP, and DHL. It provides over 60
BOD types used in 12 sample scenarios and directly supports multi-step transac-
tions, but does not include a separated BO model (BOs and BODs are not sepa-
rated). The GS1 organization cooperation with UN/CEFACT, butdue to its
nearness to EANCOM-only uses UN/CCTS methodology, not the UN/CCL com-
ponents (e.g., [38] shows the difference in address formats). Instances are freely
extensible (by so-called extensions), but the schema can only be extended by the
GS1 organization following a formal process.
UBL (Universal Business Language) [36] from OASIS is a generic, open, and
free, cross-domain, XML-based interchange format for business documents, which
can be expanded to meet the needs of specic industries. UBL is based directly on
UN/CEFACT CCTS, whose core components form the basis of 31 (in version 2.0)
or 64 (version 2.1) business documents, thus it is less powerful than OAGIS
especially for logistics. UBL was developed with the aim, among other things, to
facilitate the exchange of data between companies belonging to different industries
and therefore cannot use a common standard.
OAGIS (Open Applications Group Integration Specication) [15] from the
OAGi is an international cross-domain transaction standard for B2B and A2A and
exists since 1996 (only its rst versions were not XML-based), it is used by over 38
industries in 89 states (05/2011); main stakeholders are IBM, Oracle, DHL, SAP,
Microsoft. OAGIS 9.5.1 consists of 84 BOs, used in over 530 BODs (including
master data exchange), the BODs are used in 64 sample scenarios, and OAGi
provides Web service denitions. One of its explicit objectives is to provide a
canonical business object model [27]. It integrates many other standards: UN/
CEFACT, ISO, OASIS, CCTS/CCL and many more and can be used together with
ebXML and is EDIFACT-compatible. The OAGi quickly adopts modern trends,
e.g. soon the JSON exchange format will be supported additionally to XML to
better support mobile devices, and there are cloud and BPMN initiatives. Most
important for us is its openness and schema extensibility by XSD overlays, in
addition to instance extensions by freely usable so-called user areas.
GDTs (Global Data Types) [44, 45] from the SAP AG is a proprietary and not
freely usable business object model also based on CCTS. SAP has successfully
shown that a global business object model stored in its Enterprise Service Repos-
itory is very benecial for tool and data integration: the model is used by many of
its applications and components.

5.2 Bridging the Business-IT Gap

Bridging the gap between business and IT is not a new topic: one approach is
composite applications [9], each consisting of a stable service layer and a rather
flexible business process layer, which denes the application. They connect the IT
and business layer for each application.
Seamless Interoperability in Logistics: Narrowing 113

Another new trend is smart process applications [46] and smart process apps
platforms, a new generation of BPMS (business process management suites). Most
big BPM providers have it on their agenda or provide rst products, evaluated by
Forrester in [47]. The borders between apps and BPM system become fuzzy, since the
apps more and more use functionality of the BPM platform. These apps are highly
flexible, loosely structured, easily usable and mainly intended for human interaction
by business people (Designed for People, Build for Change, Critical to Success).
In [48] some important underlying techniques, namely front-loading and look-
ahead are introduces to minimize the business-IT gap for BOs (data), and for
services and processes (behavior), therefore we summarize it here:
Front-Loading means that the domain user provides additional information that
he knows and that would otherwise be missing for the implementation. We instead
use pre-implementations that may be provided by IT experts of Logistics Mall
service providers, not implementations derived automatically from front-loading
information. Front-Loading includes aspects like organizational units and roles,
BOs, domain-specic services information, flexibility points (where e.g. an
exchange of services is expected), foreseeable exceptions, form specications (to
derive GUIs), etc.
Look-Ahead means to nd and reuse building blocks of all kinds (which already
have those Front-Loading aspects), i.e. pre-implemented building blocks for data,
behavior, exceptions and more. That is our focus, since the Logistics Mall business
model is a shop of pre-implemented building blocks, combinable with correctness-
by-construction in mind.
BOs (Data): Executable process models need specications of all data objects
(or BOs) to precisely dene the interface of the process model and its internal data
flow, i.e. for starting process instances, calling services and for user interaction via
GUIs. In process modelling normally only the name and the most relevant attributes
of data types are specied. Using standardized BOs, all additional implementation
information is already available.
Front-Loading: Only domain experts have the knowledge of required domain
information and its structure, formats and codes. Using standard BOs, the data
analysis has already been done by several domain experts.
Look-Ahead: Existing data object types should be reused when designing
business process models. Using the BO model, there already exist detailed
specications (documented UML class diagrams and structured data from the
BO Metadata Repository) and even implementations (Java classes and the BO
Instance Repository). To enable reuse of existing BO types through look-ahead,
these are easily accessible to business process designers via the LPD tool [28].
As an advantage, efforts for designing, specifying, and implementing BOs are
signicantly reduced [12]. Even more important, business process models then
refer to the same data types (BOs) as existing services and human tasks. Due to
this harmonization, the overall efforts for implementing the transformation of a
data structure dened at business level into another one used at technical level
can be avoided [12] during service calls and process execution.
114 M. Bhmer et al.

Services (Behavior): Several activities in a process model use services (or apps/
applications) for their implementation. During runtime, process instances have to
communicate with these services via exchange of data objects and events, the latter
in asynchronous situations. Thanks to BPMN 2.0, there is no more need to bridging
the gap between business process models and system process models or service
composition specications like BPEL, as described in [49], this narrows the busi-
ness-IT gap already signicantly.
Front-Loading: In general, for each service to be used in a process model, its
functionality and the BOs of its input and output parameters need to be pre-
specied, and which IT system shall offer this service. Using the LPD, a business
process modeler may then use this information. Different systems may store
different attributes for BOs, i.e. those they need. To foster front-loading, the
Logistics Mall MMP [33] supports describing offered apps in business view that is
pre-linked to the technical descriptions needed for service execution, and service
governance is dened. Technical descriptions like WSDLs and XSD data type
specications are used, but they are not comprehensible to business process
designers having no or only limited IT skills and are therefore hidden from them.
Look-Ahead: Usually, service descriptions and service operations are published
in a repository (SOA repository or LDAP in our case). To enable look-ahead, a
service should be described in a language (or graphical notation) easily
understandable to business process designers [48], as described above. Fur-
thermore, it must be easy to search for needed services and to access and
understand related descriptions. Based on this, business process designers can
nd appropriate services and use them in corresponding process steps. Existing
technical service specications are pre-linked with these business level speci-
cations to avoid unnecessary implementation steps. Reuse of existing services
reduces efforts and costs for service implementation or service renting. As a
disadvantage, adjustments of the dened process logic to the available service
set might become necessary.

6 Conclusion

We presented the new Logistics Business Objects model BO4L developed for and
used by the Logistics Mallits objectives, design, development methods, supporting
software and how it lls the business-IT gap.
BO4L is a Logistics domain model developed by a team of Logistics experts as
an abstraction of a canonical, flexible, modern, and lively technical standard,
namely OAGIS. By this construction, the domain model is highly standards-based
and independent of specic applications and good acceptance in the Logistics
domain is predictable, becoming the new oil of logistics.
By supporting a simple business view and a tightly coupled complex standards-
based technical view, this standard reduces horizontal gaps in two dimensions: it is
Seamless Interoperability in Logistics: Narrowing 115

a common language for Logistics experts of different cooperating parties and a

common technical transaction standard for IT systems of all partners, and it is used
for tool integration of Logistics Mall offers.
Additionally, by its construction as an elaborated abstraction of a technical
standard, the Logistics BO model is also capable to narrow the vertical business-IT
gap, both on data level and on activity level. The latter is used for a new paradigm
in business process modelling.
The Logistics BO model is not only a data model with extensive documentation,
but also consists of many technical artefacts, i.e. different technical representations
of all business objects and also of software components supporting the handling of
business objects at design time and at runtime.
We only sketched the main ideas of app integration, standard-based app com-
munication, and app orchestration by process models, all with a focus on business
objects. For a detailed discussion on how applications or apps get Mall-conformant
BO-based interfaces or how process models are modelled and executed in the
Logistics Mall we refer to [33] and [28] respectively.

References

1. Fraunhofer Gesellschaft: Business objects for logistics. http://www.bo4l.org (2013)

2. EfzienzCluster LogistikRuhr. http://www.efzienzcluster.de/en
3. Fraunhofer Gesellschaft: Logistics Mall. http://www.ccl.fraunhofer.de/en.html (2013)
4. Holtkamp, B., Steinbuss, S., Gsell, H., Loeffeler, T., Springer, U.: Towards a logistics cloud.
In: 2010 Sixth International Conference on Semantics, Knowledge and Grids. http://
ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=05663532 (2010)
5. Bundesministerium fr Wirtschaft und Technologie: Exponat IKT-Innovationen in der
Logistik Wachstumsmotor fr Unternehmen, Nationaler IT Gipfel. http://www.post-und-
telekommunikation.de/PuT/1Fundus/Dokumente/7._Nationaler_IT-Gipfel_2012_Essen/20_it-
gipfel-2012-exponat-ikt-innovationen-in-der-logistik,property=pdf,bereich=itgipfel,sprache=
de,rwb=true.pdf (2012) (in German)
6. Nationaler IT Gipfel. http://www.it-gipfel.de
7. Hinssen, P. (ed.): Information is the new oildrilling new sources of innovation, across
technology. http://datascienceseries.com/assets/blog/GREENPLUM_Information_is_the_new_
oil-LR.pdf
8. Nettstrter, A., Geien, T., Witthaut, M., Ebel, D., Schoneboom, J.: Logistics Software
Systems and Functions: An Overview of ERP, WMS, TMS and SCM Systems. Springer,
Heidelberg (2013)
9. Stiehl, V.: Composite application systemssystematisches konstruieren von
verbundanwendungen unter verwendung von BPMN, dissertation, TU darmstadt. http://
tuprints.ulb.tu-darmstadt.de/2751/ (2011) (in German)
10. DIN German Institute for Standardization e.V.: Economic benets of standardization
Summary of resultsFinal report and practical examples, Beuth Verlag. http://www.din.de/
sixcms_upload/media/2896/economic_benets_standardization.pdf (2000)
11. Berlecon research: E-business standards in Germanycurrent state, problems, prospects.
http://www.berlecon.de/studien/downloads/Berlecon_eStandards2010_ExecSum.pdf (2010)
12. Bhmer, M., Daniluk, D., Schmidt, M., Gsell, H.: Business object model for realization of
individual business processes in the logistics domain. In: Clausen, U., ten Hompel, M.,
Klumpp, M. (eds.) Efciency and Logistics, pp. 237244. Springer, Berlin (2013)
116 M. Bhmer et al.

13. UN/CEFACT: Core components technical specication, version 3.0. http://www.unece.org/

leadmin/DAM/cefact/codesfortrade/CCTS/CCTS-Version3.pdf (2009)
14. UN/CEFACT. http://www.unece.org/cefact.html
15. Open Applications Group, Inc.: OAGIS release 9.5.1. www.oagi.org (2012)
16. Schmidt, M., Tilp, M., Bhmer, M.: Fachliche dokumentation der geschftsobjekte fr die
logistik, release 1.2.2 blaze, technical report, fraunhofer gesellschaft (2013)
17. Scott-Sabic, V.: Dictionary of Logistics. Deutsch-Englisch/Englisch-Deutsch. Springer,
Heidelberg (2004)
18. Heidenblut, V., ten Hompel, M.: Taschenlexikon Logistik. Springer, Heidelberg (2011). (in
German)
19. GS1: GS1 General Specications. http://www.gs1.org/docs/gsmp/barcodes/GS1_General_
Specications.pdf (2014)
20. UN/CEFACT: Introducing UN/EDIFACT. http://www.unece.org/trade/untdid/welcome.html
21. Open Applications Group, Inc.: Business object document (BOD) message architecture for
OAGIS release 9. + , OAGi white paper, version 1.0. http://www.oagi.org/oagi/downloads/
ResourceDownloads/2011_0408_BOD_Message_Architecture_V9x.pdf (2011)
22. GS1: GS1 XML. http://www.gs1.org/ecom/xml
23. Fraunhofer IAO: openTRANS. http://www.opentrans.de
24. Open Applications Group, Inc.: Users guide for extending OAGIS 8, OAGi reference guide,
version 1.0. http://www.oagi.org/oagi/downloads/ResourceDownloads/Extending_OAGIS_8.
pdf (2003)
25. Open Applications Group, Inc.: Extending OAGIS 9.x through overlaysOAGi cheat sheet.
http://www.oagi.org/oagi/downloads/ResourceDownloads/Extending_OAGIS_v9.doc (2009)
26. Open Applications Group, Inc.: OAGi development methodology, OAGi white paper.
http://www.oagi.org/oagi/downloads/developmentmethodology/2013_OAGi_Development
Methodology.pdf (2013)
27. Open Applications Group, Inc.: OAGISA Canonical business language providing both
vertical and horizontal requirements, OAGI white paper, version 1.0. http://xml.coverpages.
org/OAGIS-Canonical.pdf (2002)
28. Bochon, I., Ivens, V., Nagel, R.: Challenges of Cloud Business Process Management.
Springer, Heidelberg (2013)
29. Object Management Group, Inc.: Business process model and notation (BPMN), version 2.0.
http://www.omg.org/spec/BPMN/2.0 (2011)
30. Steinbu, S., Weienberg, N.: Service design and process design for the logistics mall cloud.
In: Yang, X., Lu, L. (eds.): Principles, Methodologies, and Service-Oriented Approaches for
Cloud Computing. pp. 185205, IGI Global, Hershey (2013)
31. Carr, N.G.: IT doesnt matter. Harvard Bus. Rev. http://hbr.org/2003/05/it-doesnt-matter/ar/1
(2003)
32. Date, C.J.: An Introduction to Database Systems. Addison-Wesley Publishing Company,
Englewood Cliffs (1977)
33. Eggemann, J., Leveling, J., Wei, N.: Business Apps Meet the Challenge of Covering
Continually Changing Logistics Requirements. Springer, Heidelberg (2013)
34. OASIS: ebXML. http://www.ebxml.org
35. RosettaNet. http://www.rosettanet.org
36. OASIS: Universal business language TC. https://www.oasis-open.org/committees/ubl/faq.php
37. Liegl, P., Zapletal, M., Pichler, C., Strommer, M.: State-of-the-art in business document
standards, industrial informatics, pp.234241. http://publik.tuwien.ac.at/les/PubDat_186928.
pdf (2010)
38. Kabak, Y., Dogac, A.: A survey and analysis of electronic business document standards. ACM
Comput. Surv. 42(3), 131 (2010). https://lists.oasis-open.org/archives/set/200909/pdf00000.
pdf (2010)
39. Melleri, I., Hiekkanen, K. ,Mykkanen, J.: comparing information models in XML-based e-
business standardsa quantitative analysis. In: IEEE 15th International Enterprise Distributed
Object Computing Conference, pp.225234 (2011)
Seamless Interoperability in Logistics: Narrowing 117

40. Lampathaki, F., Mouzakitis, S., Gionis, G., Charalabidis, Y., Askounis, D.: Business to
business interoperability: a current review of XML data integration standards. Comput. Stan.
Interfaces 31, 10451055 (2009)
41. Li, H.: XML and industrial standards for electronic commerce. Knowl. Inf. Syst. 2(4), 487
497 (2000)
42. UN/CEFACT: Core components library. http://www.unece.org/cefact/codesfortrade/unccl/
ccl_index.html
43. BMEcat. http://www.bmecat.org/English/index.asp
44. SAP: SAP Netweaver CE 7.2 documentation, global data types. http://help.sap.com/saphelp_
nwce72/helpdata/en/b8/870116599c48a4aa2bb15b7a9967da/frameset.htm
45. SAP: Global data types. http://wiki.sdn.sap.com/wiki/display/GDT/SAP+Global+Data+Types
+(GDTs)
46. Bartels, A., Moore, C.: Smart process applications ll a big business gap, forrester report
(2012)
47. Bartels, A., Moore, C.: Smart Process Applications, Forrester Wave (2013)
48. Bauer, T., Buchwald, S., Reichert, M.: Improving the quality and cost-effectiveness of
process-oriented, service-driven applications: techniques for enriching business process
models. In: Service-Driven Approaches to Architecture and Enterprise Integration, pp.104
134. http://dbis.eprints.uni-ulm.de/936/ (2013)
49. Buchwald, S., Bauer, T., Reichert, M.: Bridging the gap between business process models and
service composition specications. In: Lee, J., et al. (eds.) Service Life Cycle Tools and
Technologies: Methods, Trends, and Advances, pp. 124153. IGI Global, Hershey. http://
dbis.eprints.uni-ulm.de/713/1/BookChapter_BRB2011.pdf (2012)
Challenges of Cloud Business Process
Management

Iryna Bochon, Volker Ivens and Ralf Nagel

Abstract Cloud Business Process Management means more than outsourcing of

the Process Editor and the Execution Engine. The overall Process lifecycle has to be
adopted to face the challenges of Cloud Computing. In this chapter we discuss these
challenges and present our solutions to reconcile cost efciency, security, avail-
ability, agility and low administration overhead.

Keywords Business process management
Cloud computing
Logistics mall

Logistics process designer

1 Introduction

In todays world you need not buy your own vehicle if you want to travel by car.
An alternative would be to join a car sharing community. In car sharing you are not
responsible for buying and servicing your own car, you just rent a car from the car
sharing pool for the time you need it. The decision which solution suites you best is
based on different influence factors like the price list, mileage, travel behavior or the
distance to the next pick up point.
The main aspect of Cloud Computing is in analogy to car sharing the change of
view from owning an IT System to just renting and using it. Instead of giving a
denition of the term Cloud Computing we want to examine some of the key
factors behind this approach.

I. Bochon (&)
V. Ivens
R. Nagel

Fraunhofer Institute for Software and Systems Engineering ISST,
Emil-Figge-Str. 91, 44227 Dortmund, Germany
e-mail: [email protected]
V. Ivens
e-mail: [email protected]
R. Nagel
e-mail: [email protected]

Springer International Publishing Switzerland 2015 119

M. ten Hompel et al. (eds.), Cloud Computing for Logistics,
Lecture Notes in Logistics, DOI 10.1007/978-3-319-13404-8_7
120 I. Bochon et al.

Elasticity: The possibility to short-dated add or remove computing resources

(CPU, memory, storage capacity) gives the Cloud customer a flexibility which
hasnt been possible in the past by obtaining own hardware or the classical
outsourcing of IT services. This advantage for the customer is tightly connected
with the problem of a precise forecast of the computing resource demands by the
cloud operator. Different approaches may lead to overbooking problems at peak
times in a Cloud data center.
Security: The security problem populated by the mass media nowadays seems to
be overrated and applies more to free of charge Cloud services like Dropbox1 or
most Google services.2 In the case of business appliances in Europe data pro-
tection is severely regulated by German and European law. The implementation
and monitoring of these rules is essential for every serious Cloud operator.
Cost reduction: This is another factor often presented by popular publications.
In reality it is quite hard to compare the cost of classic IT management with
Cloud computing. What is the equivalent value of Cloud elasticity and flexi-
bility? Neglecting these facts is comparing apples and oranges. For the Cloud
operator it is easier to balance the utilization of the hardware between multiple
customers and to absorb performance peaks of a single customer. If peaks of all
Cloud customers appear simultaneously a real cost reduction may be
unachievable.
Remote Access: Cloud Computing resources (IaaS,3 PaaS,4 SaaS5 [1]) are
accessed through a network (not necessarily the public Internet). That means the
physical computing center behind the Cloud may be located anywhere around
the world. This potential advantage must be balanced to the fact, that the legal
restrictions over the world are quite heterogeneous. Data protection and gov-
ernmental control differ in various parts of the world [2].
Availability: In most businesses a high availability of Cloud Computing
resources is required. Any single point of failure could damage the overall
availability extremely. All involved components in Cloud Computing like
servers, storage, networking, operating systems, power supplies, cooling and so
on may lead to fatal blackouts and thus must be kept failsafe within the desired
specications.
Looking at the advantages of Cloud Computing it is obvious to think about a
combination of Cloud Computing and Business Process Management (BPM).
Business Processes are ubiquitous in our enterprise world. Sometimes they are
hidden in standard operating procedures. A better way is to use any kind of

1
http://www.dropbox.com.
2
Google Mail, Drive, Picasa and so on.
3
IaaS: Infrastructure as a Service.
4
PaaS: Platform as a Service.
5
SaaS: Software as a Service.
Challenges of Cloud Business Process Management 121

Business Process Notation like EPC,6 BPEL7 or BPMN.8 Using a Business Process
Notation for the execution and monitoring of processes is the state-of-the-art
technology in driving an enterprise. With BPMN 2.0 there exists a standardized
notation including an executable semantic. Various tools and execution engines
support this standard released in 2011 by the OMG [3].
BPMN is an evolution of other basic modeling concepts and languages like Petri
nets or state machines. There are scientic researches proving that all BPMN graphs
can also be modeled as Petri nets [4]. Those graphs do not contain execution
information, so in order to execute a Process model there has to be a special
description language for the execution. One language established for that purpose
was the Business Process Execution Language, which describes the orchestration of
Web services [5] included in a Process model. One of the weaknesses of this
language was the missing support for human activities, so some extensions of
BPEL including human tasks emerged (BPEL4People). For some time the com-
bination of BPMN for modeling and BPEL variants for execution was used until
BPMN 2.0 was nally released.
Typically domain specialists design Process models, but in most cases they do
not have the needed technical experience and understanding of the infrastructure or
services for the execution of a process. Consequently, a designed Process model
needs some modications and technical renements performed by IT specialists to
make it executable. For the execution of a Process model every activity must be
mapped to a service implementation (e.g. a Web service) in order to provide the
desired functionality. Additionally, the outputs of activities have to be wired with
the inputs of successive activities. Incoming external event have to be delivered to
corresponding Process instances. Outgoing events must be delivered to the correct
receiver.
To execute a Business Process model, there have to be several technical
renements to the composed model. Process internal and external events must be
wired correctly for the execution of the model.
An executable Business Process Notation covers the following key aspects:
Control flow: the order in which the Process model elements are executed
(Activities, Gateways, ).
Data structure: the denition of data stored in the Process context and
exchanged with services.
Data flow: describes the flow of data between Processes and services.
Messages and Events: specialization of the data structure and flow for com-
munication and synchronization between Processes and services.

6
EPC: Event-driven process chain.
7
BPEL: WS-Business Process Execution Language.
8
BPMN: Business Process Model and Notation.
122 I. Bochon et al.

User and Roles: specialization of the data structure representing the organization
chart of peoples interacting with the Process.
Application landscape: the applications providing the service implementations
for the activities.
A typical way to exchange data between Process activities is the use of a Process
context. A Process context works like shared memory for all activities in the
Business Process. Data objects can be stored in Process variables and every activity
can access or update it. Required input and output objects of a Process activity must
be available in the Process execution context in order to execute this Process.
Some typical problems with shared memory concepts are:
Data compatibility: the data exchanged between services must be compatible for
successful execution.
Context size: a growing size of Process context may lead to performance and
scalability problems.
Transactions: in concurrent Processes multiple instances or process threads may
try to alter the same data at the same time. This may lead to inconsistency in the
data and transaction methods are needed to control the modications.
Synchronization: the synchronization of data duplicated from Process context to
external data sources may also lead to data inconsistencies.

2 Cloud Business Process Management (CBPM)

Cloud Business Process Management means more than moving of a Process Editor
and an Execution Engine into a Cloud infrastructure. Although it is possible to use
IaaS or PaaS to operate a BPM infrastructure in the Cloud, virtualization of these
servers does not give a big advantage over classic IT management. Hence, in the
following we focus on SaaS based solutions.
CBPM should focus on creating (modeling) and using Processes rather than
implementing and operating the services. Therefore, we see a clear separation of
duties between the two stakeholders involved:
The Cloud operator is responsible for running and maintaining the Business
Process environment in the Cloud together with the connected SOA9 services.
This includes setting up new or updating existing services and applications.
The (Cloud) customer of CBPM designs, tests and uses their Business Processes
orchestrating the selected Cloud services.
As a consequence from this separation of duties, privileged accounts for the
underlying operating systems, databases and so on are not issued to the Cloud
customer. Even the execution of foreign or uncertied code by the Cloud operator

9
SOA: Service Oriented Architecture [6].
Challenges of Cloud Business Process Management 123

Fig. 1 Business Process

Lifecycle

may be legally problematic and dangerous. Uncertied code provided by the cus-
tomer could harm the system health of the Cloud infrastructure or contain vul-
nerabilities of any kind. Thus a Business Process designed and used in the Cloud
may only consist of predened and by the Cloud operator certied building blocks.
Designing a Business Process goes through some development steps, well
known from software development techniques (cf. Fig. 1). In the beginning a
Process is typically designed in a graphical notation. This step is iteratively per-
formed and can be accompanied by internal simulation phases, which help the
designer evaluating his newly created or altered Process model. The design phase is
usually followed by a test phase. During the test phase a Process model can be
executed. If necessary the Process model and depending artifacts rst have to be
deployed. In most Execution environments the deployment of new Process models
is simply based on XML les. More complex is the preparation of all services
referenced by a Process model. If a service hasnt been used by a Process model
before, it must be installed and congured before usage. Often the Process model
has to be updated to reflect the current IP or URL of the newly deployed service.
An execution of a Process in the test phase must be clearly marked preventing
misinterpretations of Process data and incoming or outgoing signals to external
systems. Any error found while testing may lead to another design phase. After a
successful test phase the Process changes to the productive phase and can be used
as desired. Any errors found in the productive phase will also lead to a new design
phase. At the end of its lifetime a Process model is undeployed, saving computing
resources and preventing the creation of new Process instances.10 The physical

10
Instance of the Process model.
124 I. Bochon et al.

undeployment of a Process model usually has to wait until all running Process
instances have been completed. Through the agility of this method a newer version
of a Process model may be deployed, tested or live executed, even while the
predecessor version is still in operation. Therefore Business Process Management
needs a built in Version management for Process models.
Additional support is required from a Processs services and applications in
different phases of the Business Process Lifecycle. Typically for testing purposes
special test instances of applications are used to distinguish between productive and
test data. Demanding a cost reduction in the Cloud, this duplication of systems is
undesirable. A better way is to nd a way to mark Process instances and corre-
sponding data as test data using only one service instance during test and productive
phases. This allows also merging test and productive execution at the same time.
With the lack of privileged accounts for the Cloud customer the Execution
environment and Process services have to support the monitoring of a Process
execution. This monitoring is needed in the test phase to make sure everything
works as desired. During the productive phase the monitoring is required to analyze
any problems which might occur during Process execution. While a broken Process
instance may be ignored during a test phase, in productive mode the Cloud cus-
tomer has to deal with these instances. For the handling of broken Process instances
a specialized repair function must be provided by the Cloud environment. With the
help of these functions corrupted data must be updated and a Process execution is
reset into a prepared state so that the execution can be continued.
Another challenge in CBPM is to connect the Cloud environment with external
systems. Only considering Cloud based systems will narrow the application area.
Especially in the Logistics domain a lot of physical hardware outside the Cloud has
to be included into a Business Process. Starting from barcode scanning and
printing, a Cloud process must communicate with external Enterprise Resource
Planning tools or Warehouse Management Systems. A specialized Gateway has to
deal with incoming and outgoing messages and data, assigning incoming data to the
corresponding Cloud customer and Process instance.
We can summarize the key challenges of CBPM as follows:
Elasticity and SaaS: this outlines the Cloud Computing aspect of CBPM (just
use metaphor extended to processes and connected services).
Executability of building blocks: CBPM assembles only executable and certied
building blocks. These building blocks especially support the different Business
Process Lifecycle phases with monitoring and repair functionalities.
Compatibility: the compatibility of data exchanged by building blocks must be
guaranteed.
Connectivity: with the help of a Cloud gateway [7] the communication to the
outside world can be established. The Gateway conguration is a part of the
Process denition.
Challenges of Cloud Business Process Management 125

3 Logistics Process Designer (LPD)

3.1 Architecture of the LPD

The Logistics Mall is a marketplace for Logistic services and applications [8].
Vendors of Logistic services and applications can offer their products as monolithic
applications or specialized services. This allows users of Logistic or IT services to
rent only required software parts for their enterprise. It is also possible to create
individual Logistic Process models, built from available services, and execute them
in the Cloud.
The development of the Logistics Mall was divided into three stages (cf. Fig. 2).
In the rst stage only monolithic applications were supported. Customers could
work with their booked applications (like Software as a Service) inside a tenant
specic Customer Access Framework (CAF). For the second stage the Logistics
Mall Cloud infrastructure was extended with an Enterprise Service Bus11 (ESB).
The ESB allows interoperability between the booked applications in the Cloud. To
harmonize the communication between applications of different vendors, a set of
standardized Business Objects (BOs) [10] has been developed, together with a
Repository able to store BOs. The last stage of the Logistics Mall aims to provide
support for Business Process as a Business Service (BPaaS).12 Monolithic appli-
cations from stage two are replaced by small services that can be arranged to
Process models. The specialized tool for building Business process models in the
Logistics Mall is the Logistic Process Designer (LPD).
The Logistics Process Designer is a Web based Process modeling tool which
uses an independent Business Process Notation (cf. Fig. 8). Target user of the LPD
is every domain specialist without the need for deep IT knowledge. The graphical
notation used is intuitive for beginners and easier to learn than most subsets of
BPMN. The LPD uses only pre-implemented building blocks making a subsequent
renement unnecessary. Building blocks of the LPD can be translated into a
selection of Business Process Languages like BPEL or BPMN 2.0, making it
possible to use available Open Source or commercial Business Process Execution
Engines.
As shown in Fig. 3, we separate three different levels: the top level consists of
the Process Modeling Taxonomy Editor (PMT Editor), the Logistics Process
Designer (LPD) and the Process Repository Frontend. With the PMT Editor the
operator of the Logistics Mall manages the Process Modeling Taxonomy containing
all available building blocks. With the LPD the building blocks from the taxon-
omy13 are composed into Process models. Finally, the Business Process Lifecycle

11
An Enterprise Service Bus is a standard-based integration platform that combines messaging,
web services, data transformation and intelligent routing in a highly distributed, event-driven
Service Oriented Architecture [9].
12
In Analogy to Software as a Service.
13
A taxonomy is a systematic and hierarchical classication of things [11].
126 I. Bochon et al.

Fig. 2 Logistics Mall stages 13

Fig. 3 Big picture: Process Management in the Logistics Mall

of these Processes is controlled with the Process Repository Frontend. The access to
all mentioned Web applications is granted through the integrated Logistics Mall
SSO Gateway.
Challenges of Cloud Business Process Management 127

The middle tier or Persistence level contains the two essential data sources: The
Process Modeling Taxonomy stored in a LDAP14 server and the Process Repository
which is an ordinary RDBMS.15 The Process Repository stores various Process
models of multiple tenants. Each model can be stored in different versions identied
by a simple version number. A special version is the head version which represents
the latest version of this model, saved by a user. In difference to the head version,
all other versions of a model are immutable. Only these versions can be tested or
deployed. A head version has to be tagged (with a version number), before the
Business Process Lifecycle controlled by the Process Repository Frontend is
started.
On the Execution level we nd all systems required for the execution of our
Processes. Key part in the execution is the Business Process Execution Engine. For
the Logistics Mall we use the Activiti Open Source BPMN 2.0 engine [12].
A Process model stored in the Process Repository is translated into an XML le
according to the BPMN 2.0 specications. The code generator is controlled by the
Process Repository Frontend and customized for the Activiti engine. The translation
uses three main input templates:
The Base template is built into the code generator and reflects any specics of
the target Execution Engine platform.
Every building block in the Logistics Modeling Taxonomy has a custom Exe-
cution Block template. For a Process the block templates of all used building
blocks are joined with the base template.
The values of Building Block parameters dened in the Process model are
extracted from the process description in the Process Repository and are merged
with the rest of the template.
The integrated Business Process Execution Engine can easily be replaced by
competitive engines. The code generator is prepared for customization simply
because the three template levels are exchangeable. Beside BPMN any executable
notation and suitable execution engine is feasible (e.g. a BPEL engine).
The two other systems in the Execution level are the Business Object Instance
Repository [10] and a Complex Event Processor (CEP). The BO Instance Repos-
itory is a specialized persistence mechanism for the Business Object Model of the
Logistics Mall [10]. The CEP is the event gateway for running Process models.
With the help of the CEP it is possible to interoperate with the Process engine.
Basic operations of the CEP are:
Starting a new Process instance.
Notify a waiting Process.
Stop a running Process (Cancelation).

14
LDAP: Lightweight Directory Access Protocol.
15
RDBMS: Relational Database Management System.
128 I. Bochon et al.

Additionally it is possible to extract data from a message event. With this data it
is possible to:
Create new Business Objects in the repository.
Update existing Business Objects.
Mark a Business Object as obsolete.
Finally events to the CEP can be sent from various origins. The following list
gives a brief overview of systems able to send events to the CEP:
Running Process instances.
The BO Instance Repository.
External systems (through the Logistics Mall gateway).
User interfaces.
Services or Logistics Mall Apps [13].
All events handled by a Process model have a description in the tenant-specic
part of the PMT. Therefore in the LPD events are treated like any other activity
from the taxonomy when building a Process model.
The Logistics Process Designer is a GWT [14] application using HTML 5 [15]
elements, running on an Apache Tomcat16 Web server. HTML 5 is still in devel-
opment but it is the future standard for modern Web pages, featuring many new
functions for Browsers without the need for Plug-ins. Its architecture is shown in
Fig. 4. The LPD Frontend is the client side of the LPD which displays the user
interface. The Canvas of the LPD is based on the yFiles for HTML Framework [16]
to render graphs and the GWT-DND17 library for drag and drop features. The
frontend communicates with the Persistence and Process Modeling Taxonomy
(PMT) backend services. The LPD Persistence uses JPA18 to access a Post-
greSQL19 Database. The Persistence API provides the functionality to save and load
Process models as well as versioning them. Among other data, the Process Mod-
eling Taxonomy stores meta information about the tenant-specic Process models
and administers the IDs by which the models are stored in the Persistence module.
The PMT is stored in an OpenLDAP20 server and is accessed through JNDI.21 The
PMT includes all available Logistics Mall applications and services combining it
with the informations about the applications and services of a tenant. Spring22 is
used for conguration and integration of the components Frontend, Persistence and
PMT. The Common component contains the data transfer objects for the com-
munication of the core components.

16
Apache Tomcat: http://tomcat.apache.org/.
17
GWT-DND: Drag & Drop Library for GWT: https://code.google.com/p/gwt-dnd/.
18
JPA: Java Persistence API.
19
PostgreSQL: http://www.postgresql.org/.
20
OpenLDAP: http://www.openldap.org/.
21
JNDI: Java Naming and Directory Interface.
22
Spring Framework: http://www.springsource.org/.
Challenges of Cloud Business Process Management 129

Fig. 4 LPD modules and

architecture

The LPD interface (Fig. 5) can be separated into three areas, the tenant inde-
pendent taxonomy tree (1 + 2), the tenant specic tree (3 + 4) and the canvas (5).
The canvas is the modeling area, where the Process models are designed. The tenant
independent taxonomy tree contains the Control flow elements as well as all Mall
applications and services. The Control flow elements are shown as number 1. These
elements are the building blocks to describe the control flow or events. The other
entries of this tree are Logistics Mall Applications, such as the Nelsan Shop [17] in
Version 1.1.0 with the Service Print delivery note (2).
The tenant specic taxonomy tree lists all applications the tenant has already
booked, as well as those he is planning to book in the future. The state of an
application (3) is indicated by its coloring, this gives an overview which applica-
tions a tenant needs to acquire to run all of his processes. Number 4 shows a list of
Process models the tenant has already created. They can be used as Subprocess
building blocks in other models.
The PMT is the key part of the LPD. Every building block used in a Process
model is taken from here. The PMT provides several functions such as classica-
tion, hierarchical structure and management of the available building blocks.
Additional meta information is stored for each building block. The user picks a
building block from the taxonomy and drags it onto the canvas. Afterwards the
blocks can be customized and connected with edges.
In the LPD the PMT is represented by two tree widgets.23 The PMT structure
consists of multiple root categories, which are connected by references in between.
One part of this taxonomy has to be administered by a Taxonomy administrator, the

23
A widget is a graphical component of the GUI, like a textbox or tree.
130 I. Bochon et al.

Fig. 5 LPD interface components

other part is changed programmatically during the design of a Process model. The
PMT has a tree structure which allows a deep hierarchy for the objects. Two groups
can be identied in the taxonomy, one part is tenant independent, the other is tenant
specic (Fig. 6).
The independent part consists of trees for Logistics Mall Business Objects [10],
Control flow elements, Logistics Mall applications and congurations. The appli-
cation tree contains information about the provider of an application, which ver-
sions exist and which services this application version offers. The Business Objects
required as inputs or outputs for these services are described as well. Execution
templates for multiple target platforms are stored for each activity of an application.
An Execution template is the link between the building block in the canvas and the
execution code in the Execution engine. The code generator uses Execution tem-
plates when translating a Business Process into an executable Process notation.
Template parameters are used to replace special placeholders within Execution
Challenges of Cloud Business Process Management 131

Fig. 6 Process model taxonomy

templates. They can be dened on any level in the application tree. Through
inheritance these parameter get passed down to the application services level. The
Logistics Mall Business Objects tree represents all predened Business Objects in
different versions from the Logistics Mall Business Object model which are ref-
erenced as application service inputs or outputs. The Control flow elements tree
contains several control and event building blocks (e.g. start, decision) needed to
design a Process model. These building blocks do not need corresponding service
implementation. The functionality is directly implemented by the Execution engine.
The Congurations tree stores several default settings like fallback icons, if a build
block does not have a custom icon.
The Tenant specic group is a tree for each tenant, containing the applications,
processes, users and roles of this tenant. Aside from general information about
applications and Processes, this tree also stores tenant specic meta information like
the status of his Process models (e.g. test or productive) or the technical details of
his application instances. It is possible for tenants to import their own users and
132 I. Bochon et al.

roles into this tree and therefore integrate them in the LPD. This enables tenants to
use their roles in the Process model.
The focus of the LPD on Domain specialists (e.g. Logistics) requires a special
view on the taxonomy, because the technical view with all public categories can be
too detailed and complex for that target group. The LPD offers a simplied mod-
eling view which hides many technical details without affecting the result of an
executable Business Process. The user can freely switch between these views.
These are the key features of the LPD:
Business Process Execution platform independent: LPD Processes can be exe-
cuted on different Process Execution platforms by translating the LPD model
into a State machine based Process language. After translation the Processes are
executed in the native language of the appropriate Execution platform.
Pre-implemented and extensible building blocks: Every Logistics Mall application
offers pre-implemented building blocks for the provided functions and services in
a taxonomy. The LPD uses these building blocks directly, without the need for
technical renement. Attached to each building block is an Execution template
fragment for the translation into the selected native Execution engine.
Data flow based on a standardized Business Object model: all Logistics Mall
applications use the same Business Object model, which is specialized for the
Logistics domain. The Business Object model guarantees the interoperability of
Logistics Mall applications from different vendors.
The LPD offers several other features to ease the modeling and execution:
Multilingual support: All taxonomy objects can be stored in various languages,
enabling the LPD to display information like application descriptions in those
languages.
Multi-tenant support: every tenant has access to a tenant specic tree. The
tenant-independent part is read-only.
Validation of the modeling elements: the validation of modeling elements is
possible because the taxonomy has information about all elements, references
and the inheritance relations between them.
Compatibility between services: the compatibility between services can be guar-
anteed because the interfaces of the services use the standardized Business Objects.
So it is easy to check if the output of one service ts the input of another (Fig. 7).
Accessibility to users: the technical view and the simplied modeling view allow
all kinds of users to create executable Process models.
Automatic code generation: the usage of Execution templates for activities in
combination with template parameters facilitates the automatic code generation
for any target platform.

Fig. 7 Compatibility test

Challenges of Cloud Business Process Management 133

Versioning: the versioning of Process models is supported through the taxon-

omy as a Process model can store its versions in a branch.
Flexible modeling notation: The available Process model elements of the LPD
are dened in the PMT, so new model elements can easily be added or the
properties (e.g. icons, parameters) of existing ones can be changed without the
need for LPD changes.
Optimum Performance: the tree structure of the LPD taxonomy is optimized for
read operations and a good performance.

3.2 Modeling of a Basic Example

The Screenshot (Fig. 8) shows the canvas of the LPD with a basic Process model.
This process is based on the Nelsan Scenario [17], presented on the Nationaler IT
Gipfel 2012 [18]. The following overview summarizes the used building blocks
(yellow rectangles connected with dotted lines are comments) (Table 1):

Fig. 8 Order Management Process model

134 I. Bochon et al.

Table 1 Building blocks of the order management process model

Building block Description
Order received The process starts with the arrival of a purchase order BO from the
online shop.
Transfer order This activity is an automated service to import the purchase order
into the work items of the Order Management System (OMS).
Check order manually In this human task (visualized by the checkbox in the activity icon)
the order is manually reviewed or declined.
Order cleared? The decision block evaluates the review result and follows either
the Yes or No path.
Fork If the order was approved, further activities are called
simultaneously.
Pick and delivery This subprocess handles the packing and distribution of the
ordered products.
Send certicate of dis- After the subprocess is nished, a certicate of dispatch is sent to
patch to shop the shop.
Send order conrmation The order conrmation is send to the shop.
to shop
Join After both paths are nished, they are joined in one control flow
and the process stops.
Stop Ends this process.
Cancel order If the order was refused, the activity cancellation service is called.
Send order cancellation This building block sends a refusal message to the shop and the
to shop process stops.

4 Related Work

The rapid development of IT and Cloud based architectures opens up new poten-
tials for Business processes. Especially company overlapping business Processes
give room for improvement of the scalability, costs and complexity through the
Cloud BPM combination. Many providers have identied those potentials and offer
commercial or Open Source Cloud based BPM solutions [19, 20].

4.1 Commercial Vendors for Cloud-BPM

Commercial Cloud BPM products are the majority of available tools. These usually
offer a broad lineup of functions for workflows, collaboration, simulation and
compatibility. The compatibility is realized through many adapters and connectors
for Business Objects. The LPD does not need adapters, as the standardized BO
model is used for communication to avoid incompatibilities between interfaces.
Many of those commercial products are well established on the BPM market [21]:
Inubit BPM-Suite from Bosch Software Innovations [22] directly generates
executable Processes from Process models using customizable templates for the
Challenges of Cloud Business Process Management 135

code generation, like the LPD. Inubit offers an own Process engine for the exe-
cution of Processes. Existing Processes can be imported from external BPM
systems.
Intalio offers the Intalio|Create [23] Web based platform which includes an own
Process engine. This engine can execute modeled Processes by involving prede-
ned constructors or scripts in various programming languages. Intalio distin-
guishes between two perspectives, a technical and a domain perspective.
The company IYOPRO offers a product with an identical name [24]. The product
makes it possible to run Process flow simulations as well as path visualization with
automatic validation of executability. The domain Process models can be automated
through extensions of the Workflow management. The IYOPRO BPM suite sup-
ports collaborative access of Process models and supports various Process engines,
including an in-house development, for the execution of those Process models.
The following tools support CBPM only as PaaS solutions, which is not the
focus of this article [21]: Appian BMP Suite [25], Cordys Business Operations
Platform [26, 27], Interstage Business Operations Platform from Fujitsu [28] and
TIBCO ActiveMatrix BPM with TIBCO Silver BPM [29, 30].

4.2 Open Source Products for Cloud-BPM

Available Open Source solutions in the market supply sufcient functionalities for
basic application areas whereas commercial tools outperform with additional fea-
tures and better support [31]:
Activiti BPM Platform [12] from Activiti is a good representative of this cate-
gory. This Process engine is used in the Logistics Mall Stage 3. The engine is
Javabased with a Model Repository to save Process models. This BPM platform
provides the Activiti Explorer to view running Process instances. The transition
from the domain model to the technical model is only possible with a switch of the
used Activiti components: a non-trivial technical renement of the domain model
from the Activiti-Modeler is only possible by using the Activiti Designer. The
platform supports test Process execution with the help of unit tests and the task
management. The Java code enables the customer to implement his own activities
[32].
Intalio/Create is also available in a free community version. In comparison to
the commercial version, the Intalio|Pipes component is not included. This com-
ponent orchestrates calls to other systems and components inside the Intalio Cloud.
IYOPRO has also basis version available as Open Source. The essential differ-
ences to the commercial version are the missing Process execution and the custom
rules for compliance checks.
136 I. Bochon et al.

4.3 Academic Solutions for Cloud-BPM

Only few academic solutions exist, but the growing interest in the subject opti-
mization of Business Processes with connection to the cloud can be ascertained.
They are primarily focused on innovative ideas, which are realized in the context of
a project and eventually lead to a nished product on the market. The scientic
sector provides the following solutions or unnished projects aside from the LPD.
The Institute of Databases and Information Systems from the University of Ulm
presented their Cloud based BPM Solution clavii BPM Cloud [33] on the CeBIT
2013. With the help of this solution, every user can work with personalized Process
views and make changes to the model. Those changes are immediately published to
all other users which are affected by the changes. The personalization is achieved by
abstraction of the Business Processes (depending on the eld of duty, the necessary
activities are shown and the other activities are hidden or grouped). The complexity
of Business Processes can be reduced through the selection of various views.
Furthermore, it is possible to change a graph based process visualization (BPMN
2.0) in a form based, a textual to an ADEPT24-visualization. The clavii BPM Cloud
facilitates case based Process changes at runtime. The solution offers a Process
modeler and Process execution is planned. The solution emerged from the science
project proView [34] from the University of Ulm, where the concepts of the Process
abstraction were taken from [35].
The Konstanzer Institut fr Prozesssteuerung (KIPS) from the HTWG Konstanz 25
works in cooperation with industrial partners on the science project BPM@Cloud
[36], where the combination possibilities of Cloud Computing and BPM are resear-
ched in separate labs (BPMN-Lab, Cloud-Lab and Mobile-Lab). Various Activities
which would prot from Cloud Business Processes are determined from the research
to optimize the Business Processes based on scientic aspects. KIPS has developed
the method BPM(N)Easy for that purpose, where the agile software development is
combined with the classical BPM lifecycle. This method was successfully tested
with the introduction of new Processes. The next steps are researching to what extend
this method can optimize already running Processes. For tooling the BPM platform
Xpert.ivy [37] from the Axon Active AG is used. A mobile application for a mobile
BPM, which uses Web services to communicate with a Workflow engine running in
the Cloud, is in development at the Mobile-Lab. With this application Processes can
be created, executed and monitored [19, 36, 38].

24
ADEPT: Application Development Based on Encapsulated Pre-Modeled Process Templates,
Process Management System developed from the University Ulm [39].
25
HTWG Konstanz: University of Applied Sciences.
Challenges of Cloud Business Process Management 137

5 Future Work and Conclusion

In this article we presented the challenges of Cloud Business Process Management.

A specialized CBPM tool suite can signicantly contribute to the optimization and
automation of Business processes in the Cloud. We analyzed available tool suites in
the market with the outlined requirements and found no adequate solution. Con-
sequently we have developed the Logistics Process Designer including the com-
ponents and tools shown in Fig. 3. The stage three of the Logistics Mall is our ideal
solution for the execution of Business processes in a Cloud for Logistics
enterprises.
Main aspect of the LPD is usability and hiding technical aspects: A user does not
need either IT knowledge or Process modeling skills. After a quick introduction
anybody is able to build a Process model with the LPD.
With the help of pre-implemented building blocks, IT renement has become
unnecessary. The code generator translates a Process model automatically into an
appropriate executable model for the integrated Execution engine. Technical details
as needed for communication with the outside world are transferred into the Mall
gateway. The Process Modeling Taxonomy has a central role in our approach. It
contains all available building blocks and denes the corresponding Execution
templates. The Cloud operator has to carefully plan and operate the PMT in order to
guarantee support for a wide range of use cases to be realized with the solution.
In the future we will complete the LPD with additional features. For monitoring
we will use the Process model in the canvas and highlight the current execution and
give information about the current Process context. Furthermore, a direct link to the
Business Object Instance Repository will give the chance to inspect and modify
Business Objects directly from the canvas in case of an exception situation. With a
Multi Execution engine support we give the Cloud customer the chance to select an
Execution engine most suitable for his enterprise. The code generator will translate
the Process model into the correct engine the Cloud customer has selected.

References

1. Mell, P., Grance, T.: The NIST denition of cloud computing, working paper national institute
of standards and technology (2009). http://www.nist.gov/itl/cloud/upload/cloud-def-v15.pdf
2. BSA: 2013 BSA Global Cloud COMPUTING Scorecard (2013). http://cloudscorecard.bsa.
org/2013/assets/PDFs/BSA_GlobalCloudScorecard2013.pdf. Accessed 29 April 2013
3. OMG: Business Process Model and Notation (BPMN), Version 2.0 (2011). http://www.omg.
org/spec/BPMN/2.0. Accessed 29 April 2013
4. Harmuth, N.: Business Process Model and Notation 2.0Evaluierung der Spezikation und
Ableitung einer formalen Semantik an ausgewhlten Sprachelementen auf Basis von Petri-
Netzen (2010), Diplomarbeit an der Fakultt fr Informatik der TU Dortmund (reference in
German)
5. World Wide Web Consortium (W3C): Web Services Architecture. http://www.w3.org/TR/ws-
arch/. Accessed 29 April 2013
138 I. Bochon et al.

6. Lawler J.P., Howell-Barber H.: Service-Oriented Architecture: SOA Strategy, Methodology,

and Technology. Auerbach Publications, Boca Raton (2008)
7. Holtkamp, B.: The logistics mallan it-architecture for logistisc-as-a-product. In: ten Hompel,
M., Rehof, J., Wolf, O. (eds.) Cloud Computing for Logistics, Lecture Notes in Logistics,
Springer (2015)
8. Gsell, H., Nagel R.: Application integration in the logistics mall. Paper presented at the SDPS
201217th international conference on transformative science, engineering, and business In-
novation. Berlin, Germany (2012)
9. Chappell, D.: Enterprise Service Bus: Theory in Practice, OReilly Media Inc., Sebastopol
(2004)
10. Bhmer, M., Schmidt, M., Weienberg, N.: Seamless interoperability in logistics: narrowing
the business-it gap by logistics business objects. In: ten Hompel, M., Rehof, J., Wolf, O. (eds.)
Cloud Computing for Logistics, Lecture Notes in Logistics, Springer (2015)
11. Wnek, J.: LSI-based taxonomy generation: the taxonomist system. In: Kantor, P. et al. (eds.)
ISI 2005, LNCS, vol. 3495, pp. 389394. Springer, Berlin, Heidelberg (2005)
12. Activiti BPM Platform Homepage. http://www.activiti.org
13. Eggemann, J., Leveling, J., Wei, N.: Business apps meet the challenge of covering
continually changing logistics requirements. In: ten Hompel, M., Rehof, J., Wolf, O. (eds.)
Cloud Computing for Logistics, Lecture Notes in Logistics, Springer (2015)
14. Google Developers: Google Web Toolkit. https://developers.google.com/web-toolkit/.
Accessed 29 April 2013
15. World Wide Web Consortium (W3C): HTML5 Candidate Recommendation. http://www.w3.
org/TR/2012/CR-html5-20121217/. Accessed 29 April 2013
16. yWorks: yFILES for HTML. http://www.yworks.com/de/products_yleshtml_about.html.
Accessed 29 April 2013
17. Bundesministerium fr Wirtschaft und Technologie: Exponat IKT-Innovationen in der
LogistikWachstumsmotor fr Unternehmen, Nationaler IT Gipfel 2012, Retrieved from
http://www.post-und-telekommunikation.de/PuT/1Fundus/Dokumente/7._Nationaler_IT-
Gipfel_2012_Essen/20_it-gipfel-2012-exponat-ikt-innovationen-in-der-logistik,property=pdf,
bereich=itgipfel,sprache=de,rwb=true.pdf (reference in German)
18. Nationaler IT Gipfel Homepage. http://www.it-gipfel.de
19. Mevius, M., Wiedmann, P.: BPM @Cloud-LabOpen Innovation Platform zur Entwicklung
und Evaluation eines agilen Geschftsprozessmanagements. In: HTWG Forum. Das
Forschungsmagazin der Hochschule Konstanz, pp. 6066. HTWG Konstanz (2012)
20. Mevius, M., Stephan, R., Wiedmann, P.: Innovative approach for agile bPM. In: The Fifth
International Conference on Information, Process, and Knowledge Management eKNOW,
pp. 160165. IARIA (2013)
21. Sinur, J., Hill, J.B.: Magic quadrant for business process management suites. In: Gartner RAS
Core Research Note G00205212, pp. 124. RA2 10182011 (2010)
22. Inubit BPM-Suite Homepage. http://www.bosch-si.com/technology/business-process-
management-bpm/business-process-management.html. Accessed 29 April 2013
23. Intalio|Create Homepage. http://www.intalio.com/products/create/challenges. Accessed 29
April 2013
24. IYOPRO Homepage. http://www.iyopro.de. Accessed 29 April 2013
25. Appian BPM Software Homepage. http://www.appian.com/bpm-software/bpm-suite.jsp.
Accessed 29 April 2013
26. Cordys Business Operations Platform Homepage. http://www.cordys.com/business_
operations_platform. Accessed 29 April 2013
27. Fleming M., Silverstein J.: IDC market scape: worldwide business process platforms 2011
vendor analysis. In: BuyerPulse Analysis: Business Process Management Software,
Competitive Analysis, IDC #228598, vol. 1 (2011)
28. Interstage BOP Homepage, http://www.fujitsu.com/global/services/software/interstage/
solutions/bop. Accessed 29 April 2013
Challenges of Cloud Business Process Management 139

29. Craggs, S.: Comparing BPM from Pegasystems, IBM and TIBCO. Taking a high-level look at
BPM solutions from three leading vendors. In: Whitepaper, 1.00 (2011)
30. TIBCO Software Homepage. http://www.tibco.de. Accessed 29 April 2013
31. Rcker, B.: Open Source BPM. Ein berblick ber aktuelle Tools und Strmungen. In:
Javamagazin, DevOps, pp. 8695 (2012) (reference in German)
32. Preusker, N., Rcker, B.: Activiti 5. Open source BPM on the next level. In: Javamagazin,
Social Media API, pp. 106110 (2010)
33. Clavii BMP Cloud Homepage. http://clavii.com. Accessed 29 April 2013
34. ProView (Personalized and Updatable Process Visualizations) Homepage. http://www.uni-
ulm.de/in/iui-dbis/forschung/projekte/proview-project.html. Accessed 29 April 2013
35. Bingmann, A.: Ulmer Informatiker auf der CeBIT 2013: Neue Anwendung garantiert
Durchblick bei komplexen Geschftsprozessen. News. http://www.uni-ulm.de/home/news-
details/article/ulmer-informatiker-auf-der-cebit-2013-neue-anwendung-garantiert-durchblick-
bei-komplexen-geschaefts.html (reference in German)
36. BPM@Cloud Forschungsprojekt Homepage. http://bpmcloud.in.htwg-konstanz.de/
BpmCloud/index.php. Accessed 29 April 2013
37. Xpert.ivy Homepage. http://www.xpertline.ch/en/xpertivy.html. Accessed 29 April 2013
38. Collaboration in der Cloud 3.0Business Process Management und Social Media. In:
Midrange Magazin, ITP Verlag GmbH, online version: (2013). http://www.midrangemagazin.
de/mm/artikel.html?id=7992. Accessed 29 April 2013
39. Bauer, T., Dadam, P.: lfcient distributed workflow management based on variable server
assignments. In: Wangler, B., Bergman, L. (eds.): CAiSE 2000, LNCS vol. 1789, pp. 94109.
Springer, Berlin, Heidelberg (2000)