Online Information Review

Architectural support for business intelligence: a push-pull mechanism

Jing‐Shiuan Hua, Shi‐Ming Huang, David C. Yen,
Article information:
To cite this document:
Jing‐Shiuan Hua, Shi‐Ming Huang, David C. Yen, (2012) "Architectural support for business
intelligence: a push‐pull mechanism", Online Information Review, Vol. 36 Issue: 1, pp.52-71, https://
doi.org/10.1108/14684521211206962
Permanent link to this document:
https://doi.org/10.1108/14684521211206962
Downloaded by JAZAN UNIVERSITY At 01:50 15 August 2017 (PT)

Downloaded on: 15 August 2017, At: 01:50 (PT)

References: this document contains references to 32 other documents.
To copy this document: [email protected]
The fulltext of this document has been downloaded 1610 times since 2012*
Users who downloaded this article also downloaded:
(2014),"A survey on recent research in business intelligence", Journal of Enterprise Information
Management, Vol. 27 Iss 6 pp. 831-866 <a href="https://doi.org/10.1108/JEIM-06-2013-0029">https://
doi.org/10.1108/JEIM-06-2013-0029</a>
(2014),"Review of business intelligence through data analysis", Benchmarking: An International Journal,
Vol. 21 Iss 2 pp. 300-311 <a href="https://doi.org/10.1108/BIJ-08-2012-0050">https://doi.org/10.1108/
BIJ-08-2012-0050</a>

Access to this document was granted through an Emerald subscription provided by emerald-srm:556029 []
For Authors
If you would like to write for this, or any other Emerald publication, then please use our Emerald for
Authors service information about how to choose which publication to write for and submission guidelines
are available for all. Please visit www.emeraldinsight.com/authors for more information.
About Emerald www.emeraldinsight.com
Emerald is a global publisher linking research and practice to the benefit of society. The company
manages a portfolio of more than 290 journals and over 2,350 books and book series volumes, as well as
providing an extensive range of online products and additional customer resources and services.
Emerald is both COUNTER 4 and TRANSFER compliant. The organization is a partner of the Committee
on Publication Ethics (COPE) and also works with Portico and the LOCKSS initiative for digital archive
preservation.

*Related content and download information correct at time of download.

 The current issue and full text archive of this journal is available at
www.emeraldinsight.com/1468-4527.htm

OIR
36,1 Architectural support for business
intelligence: a push-pull
mechanism
52
Jing-Shiuan Hua
Department of Information Management, National Chung-Cheng University,
Received 23 June 2010
Accepted 1 May 2011 Chia-Yi, Taiwan
Shi-Ming Huang
Department of Accounting & Information Technology,
National Chung-Cheng University, Chia-Yi, Taiwan, and
Downloaded by JAZAN UNIVERSITY At 01:50 15 August 2017 (PT)

David C. Yen
Department of DSC & MIS, Miami University, Oxford, Ohio, USA

Abstract
Purpose – As business globalisation and internet usage continue to grow, the internet-based version
of data warehouse systems (DWS) is expected to improve traditional DWS. However applying the
web-based interfaces to client-server-based DWS structures may cause problems such as inflexibility,
inefficiency, loss of scalability, and threats to security. These arise due to the complexity of
manipulation and management of heterogeneous data with various categories of decisional tasks. This
paper seeks to develop a flexible mechanism by applying Extensible Markup Language as a
foundation for an internet-based DWS and to overcome the weaknesses of solely client-server-based
DWS architecture.
Design/methodology/approach – For better control and security the proposed architecture utilises
an embedded pull-push mechanism to propagate the distributed decision information. This research
also justifies the feasibility of the proposed mechanism by implementing a prototype, evaluating its
performance, and conducting a real business case study.
Findings – The results indicate that the mechanism can not only improve DWS scalability and
efficiency, but also enhance security.
Originality/value – The proposed architecture provides a support mechanism for business
intelligence to efficiently and flexibly help companies make the right decisions in real time, grasp
business opportunities and gain competitive advantage.
Keywords Data warehouse, Data cube, Extensible Markup Language (XML), Business intelligence,
Push and pull mechanism, Star schema, Data handling, Database management systems
Paper type Research paper

Introduction
As the popularity of the internet grows, it is widely recognised that information
available on it is an extremely important external resource. The internet is actually a
rich repository for the information resources that enable people to solve problems by
Online Information Review
Vol. 36 No. 1, 2012
pp. 52-71 The research has been supported by the National Science Council, Taiwan, ROC, under Grant
q Emerald Group Publishing Limited
1468-4527
No. 97-2410-H-194-074-MY3. The authors of this research appreciate their financial support and
DOI 10.1108/14684521211206962 encouragement.
 applying the results obtained from selective and heuristic searches. Business Support for
intelligence has been defined as solutions applying information technologies to retrieve business
heterogeneous and distributed resources in order to interpret, categorise, and integrate
them, and then to formulate any potentially usable knowledge by employing analysis intelligence
mechanisms (Vine, 2000). However knowledge is a result of a rational process of
critically understanding data and information in their contexts (Will, 2006). In today’s
world data are so numerous that technology is needed to cope with this knowledge 53
resource.
Data warehouse systems (DWS), one of the dominant components of business
intelligence, have been widely adopted as an extension of database systems developed
to support decisions. The evolution of DWS has occurred mainly due to the
accumulated demands of exchanging data from various resources in order to better
utilise them for decision making. Policymakers often require various dimensions of
Downloaded by JAZAN UNIVERSITY At 01:50 15 August 2017 (PT)

data in order to generate a quick response to change. To satisfy this need DWS utilise a
multidimensional data type, data cube (DC), in support of today’s online analytical
processing (OLAP) tasks (Huang et al., 2007).
Although DWS are widely adopted by enterprises as one category of
decision-making support tools, there are some problems, including (Mehedinţu et al.,
2008):
.
Topological compatibility and cost. The two-tier client-server-based infrastructure
makes the system design and maintenance difficult and costly. Information can
only be browsed by using designated client software. Moreover scalability
bottlenecks occur, particularly in cases where fast-client solutions are provided
while handling complex reporting requirements and large numbers of users.
.
Decision support within virtual organisations. The traditional client-server-based
system is considered inadequate for meeting the demand of virtual organisations
to link external suppliers with customers. The success of a virtual organisation
depends heavily on whether or not the organisation can cooperate with its
partners and gain its customers’ confidence. Consequently restricting
information access to a small number of highly trained specialists within an
organisation is no longer considered sufficient. More information-sharing
capability is needed among external partners, including various suppliers and
customers.

The delivery of data warehouse information to decision makers throughout the

enterprise has been an expensive and challenging task. In the most exciting and
fastest-growing fields such as the internet/intranet and the world wide web, partial
solutions are available to remedy these aforementioned issues because it is possible to
apply the associated web technology to traditional data warehousing, which can result
in such benefits as improved cost savings and productivity (Mehedinţu et al., 2008).
Therefore many organisations expect a web-based DWS to improve the efficiency of
decision making, information sharing, and service personalisation (Grabova et al.,
2010). However a number of obstacles still exist in the implementation of web-based
DWS in general, and in the construction and utilisation of a distributed
decision-supporting mechanism in particular (Moeller, 2008). These obstacles include
the following:
 OIR .
Flexibility. In terms of remote access, it is difficult to identify the number of users
36,1 who will be accessing the DWS for various decision-related tasks, especially
those with different requirements.
. Efficiency and scalability. In a DWS environment the complexity of user queries
and the large amount of distributed data involved in the queries add an extra
burden, such as costs of communication, ad hoc queries and aggregation, data
54 transformation, results display, and more complex processing to facilitate
scalability.
.
Security and personalisation: In the case of personalised and distributed
decision-supporting scenarios, security considerations involve not only managing
authority and granting privileges to different users on the systems, but also
preventing theft or other harmful operations from affecting the reliability and
dependability of information (McFadzean et al., 2007). Personalisation refers to
Downloaded by JAZAN UNIVERSITY At 01:50 15 August 2017 (PT)

what users ask for (tasks) and what they get in return (data/information).

Since Extensible Markup Language (XML) provides flexibility, self-description,

self-definition, and ease of use, it is used to define a new DWS architecture (Ravat
et al., 2010). We propose an architectural solution for web-based DWS construction with
XML structural metadata to overcome the drawbacks of current internet-based DWS. In
such architecture dimensional information will be presented with open XML expressions
in order to overcome the complexity problem associated with heterogeneous analytical
data integration among DWS and their data resources. In addition with the “pull” and
“push” operations as fundamental elements of personalised decision supporting
functionalities, users will be able to access the information directly related to specific
decision-making scenarios in a routine or on-demand manner. In routine situations it is
assumed that designers have resolved any semantic and pragmatic ambiguity. In ad hoc
situations the users are responsible for resolving any semantic and pragmatic ambiguity.

Literature review
Requirements for strategic decision support facilities
For centuries managers and executives have used available knowledge to make
decisions that influence various functional layers ranging from routine operational and
managerial activities to strategic tasks within an enterprise. The decision problems in
the strategic layer are, by nature, more ad hoc and qualitative than routine and
quantitative. In addition the resources required for strategic decision support can be
available either internally or retrievable externally. Therefore the design and
implementation of decision support facilities for strategic purposes is indeed the most
complicated task (Ariyachandra and Watson, 2010).
Many prior studies have focused on the deployment and utilisation of decision
facilities with strategic purposes. Some of them discussed the feasibility of developing
unified interfaces, integrated contents, and personalised functionalities, since they are
a technology for information delivery to all business end-users in a collaborative or
even distributed manner (Cheung and Babin, 2006). The potential of the web is
considerable in this context. Consequently a web-based decision tool is indeed a
potential solution. We attempt to investigate the underlying decision mechanisms in an
enterprise in general and in a virtual organisation in particular by applying web-based
decisional tools.
 Applying multidimensional data models for data warehousing Support for
DWS have made it possible to extract and store data temporarily in a distributed data business
mart (the access layer of the data warehouse environment that is used to get data out to
the users) for business intelligence purposes. Such data can be properly analysed, intelligence
summarised, and finally applied subjectively in decision-making activities with decision
support systems (Moeller, 2008). The easiest way to meet such requirements is to create a
number of multidimensional data marts corresponding to a specific, dependent, and 55
subjective decision-making process from the data warehouses in an enterprise. To
construct multidimensional DWS, DC have been widely applied as an effective data
modelling approach to DWS. Subjective decisional data summarisation from different
dimensions and with various constraints can be generated by multidimensional
aggregation operations on the data sources of DC (Kumar et al., 2008).
OLAP operations with DWS mainly focus on query optimisation based on underlying
Downloaded by JAZAN UNIVERSITY At 01:50 15 August 2017 (PT)

multidimensional data models (Chung et al., 2007). Within the previous research in the
DC area, the heterogeneity and distributive capability of DC in a virtual organisation are
widely investigated issues (Eavis, 2009; Furtado, 2009; Wu et al., 2009). It is therefore
recognised that multidimensional DC are the necessary foundations for building a data
warehouse on the subject of distributing such decisional data resources among partners
or individuals within a virtual or networked organisation.

Web-based platform for distribution of decisional information

It has been recognised that the web is the dominant platform for data publishing and
sharing tasks (Bhargava et al., 2007). Transparency of decisional information among
partners can therefore be better achieved due to the underlying web-based open
standard. Consequently web-based decision-making tools, including web-based DWS,
can potentially enhance distributed decision supporting activities with user-friendly
unified protocols and standardised interfaces. However yet another problem with
current DWS is the implementation of personalised decisional experience and
information from various decision makers in a distributed organisation. Since users’
requirements are inherently time-related and subject-variant, it is difficult to optimise
service provided by current DWS (Moeller, 2008). It would be costly and
time-consuming to construct unified DWS that satisfy all potential users with
current technology. We therefore propose a personalised and distributed decision
process to be achieved by improving the architecture of a web-based DWS with
embedded user-pull (on demand) and technical-push (automatic) mechanisms.

XML-based warehouse for heterogeneous decisional information

To design improved web-based DWS, XML is used as a meta-language to define data
semantically (Møller and Schwartzbach, 2006; Tummarello et al., 2008). With respect to
content management XML is a platform-independent standard which allows
information in XML-based documents to freely travel around a distributed and
heterogeneous information system environment at reduced cost (El-Gayar and
Tandekar, 2007). Consequently XML-based warehouse concepts have been emerging
during the past decade. In this field prior research proposing some typical approaches,
such as Xyleme, XCube, and DAWAX, can be classified into three categories: XML
web warehouses, XML data warehouses, and XML document warehouses (Mahboubi
et al., 2009).
 OIR Although a lot of concepts regarding XML warehouse models or web systems can
36,1 be found in the previous literature, insufficient emphasis is given to standard XML
warehouse architecture seamlessly combining these emerging concepts related to
XML-based warehouses. This paper proposes an integrated architecture covering
these three kinds of concepts from past studies and which can meet the requirements of
flexible business intelligence on the internet. In previous research the authors proposed
56 a methodology to develop an XML meta-model for multidimensional DC (Huang and
Su, 2002). Based on these findings the integrated architecture for an XML web
warehouse has now been built. It improves the functionality of traditional web-based
warehouses and supports the transformation of XML documents to facilitate frequent
user-query operations. Furthermore a push-and-pull strategy is pursued in the
integrated architecture. Personalised decision-making information can therefore be
properly modelled and securely delivered in different scenarios.
Downloaded by JAZAN UNIVERSITY At 01:50 15 August 2017 (PT)

Internet-based solutions with a push-and-pull mechanism

Architectural overview
Figure 1 shows the architecture of our XML-based DWS on the internet. A three-layer
topology is used:
(1) Data warehouse service. DWS and analysis services are encapsulated as
back-end data sources for an internet-based DWS in order to manage sensitive
DC for personalised use and security.
(2) Presentation transformation service. This service is responsible for
transforming the DC from an internal data format to an open XML-based
expression in order to improve sharing capability and understandability from
both the users’ and applications’ perspectives.
(3) User support service. A security mechanism is required by an internet-based
DWS, since data provided by the DWS is strategically sensitive to users and
enterprises. Enterprises will not allow data provision and analysis facilities to
be utilised by any unauthorised users, including competitors and spies.
Therefore the push-and-pull mechanism is recommended to provide a

Figure 1.
XML-based data
warehouse system on the
internet
 personalised decision-making experience with ad hoc decisional resources. Support for
Technically the service plays the role of a controlling gateway for authorised business
communication and resource coordination.
intelligence
Re-description of original data for analytical purposes
The first task of an internet-based DWS design is the re-description of the original data
in an XML-based data format. To create an XML-based DC the first step is to 57
transform DC metadata into an equivalent XML cube schema according to a previously
proposed methodology (Huang and Su, 2002). Then the original object data will be
transformed into an XML-based DC with metadata according to a corresponding
XML-based DC schema. Since a user’s browsing facility is required for a deliverable
DC, wrapping a browsable XML-based DC into an Extensible Style Language
Template (XSLT) is necessary for the representation of XML-based documents (W3C,
Downloaded by JAZAN UNIVERSITY At 01:50 15 August 2017 (PT)

2007).
XML-based meta model template for data cube. As discussed above a meta-model
template is required to transform the data in a browsable DC. This is illustrated as an
entity relationship diagram in Figure 2. The elements that constitute a DC and the
relations among them are regarded as entities and relationships of the entity
relationship diagram, respectively. Relationships between a DC table entity and related
fact/dimension table entities are illustrated. The interesting data properties in a fact
table are identified as fact attributes within a separate entity. Likewise dimension
properties are regarded as dimension attributes within a separate entity. Various
contexts can be represented by dimensions to analyse facts by mapping fact attribute
entities and dimension attribute entities.
XML-based cube schema transformation. The first step in transforming an original
DC to an equivalent XML-based document is to investigate the semantics of the cube,

Figure 2.
XML-based meta-model
template for a dimensional
data cube
 OIR which can be expressed in a well-known star schema. The semantic data model can be
36,1 transformed into an XML-based document according to the meta-model template
shown in Figure 2. It is used to apply XML documents as an expression and an XML
multidimensional model of both schema and content information. The metadata can
then be modelled as an XML document according to the meta-model.
For nonvolatile or routine DC the extraction is performed automatically when the
58 DC has been defined. It must be stored in the repository for efficiency’s sake, since it is
a routine operation that is stable by nature. Conversely for temporary dimensional
query operations, the corresponding XML-based schema information is generated on
demand. It is optional for systems and users to store such information in the metadata
repository.
Figure 3 provides an example of an XML multidimensional model associated with the
star schema model of the sales cube. The sales cube corresponds to a set of XML facts
Downloaded by JAZAN UNIVERSITY At 01:50 15 August 2017 (PT)

categorised by dimensions. One XML document helps store information (including

dimension references and numeric measures) related to a set of facts. The dimension
reference “dim_id” is similar to the indexes used to link dimension data with fact data
and numeric measures derived from the records in the fact table. Each dimension datum
is described in one XML document containing hierarchical information, attributes, and
instance values of the levels. Rollup attributes are the level references used to link a
dimension level and its instances to the next level in an aggregate manner.
XML-based cube data transformation. The component module in the presentation
transformation service is responsible for the generation of an XML-based DC document

Figure 3.
An example of an XML
multidimensional model
 from the DC in the back-end DWS, according to the content of an XML-based schema Support for
document either routinely or on demand. For users a five-step procedure to create an business
XML-based DC from data warehouse operations is required.
(1) Personalised requirement definition. The first step provides a requirement
intelligence
definition based on user requests. User requests include not only the condition
queries created according to a pre-defined decision policy for routine
information but also the manual queries from various types of decision users 59
with ad hoc information. Therefore any ambiguity of user requests incurred
from the semantic and pragmatic requirements should be avoided by designers
in a routine situation and by users in an ad hoc situation.
(2) OLAP operation. The requirement of user requests must then be transformed to
corresponding OLAP-aggregated query scripts in order to retrieve the results
from a back-end DWS service. These scripts are used to prepare the DC and
Downloaded by JAZAN UNIVERSITY At 01:50 15 August 2017 (PT)

specify the data set that provides the information for each analytic dimension of
DC requested by the query. Next these scripts will execute the query and then
return the result data set for the DC when the query is completed and the
obtained results from the query are not empty.
(3) Result data transformation according to pre-generated star schema. When the
non-empty result data set for some DC is returned from the back-end DWS, then
cube data will be transformed into an XML-based format as an XML-based
multidimensional DC according to pre-generated XML-based schema
information stored in the schema repository. The generated DC is either a
directly browsable object appearing on client viewers or a systematically
parsable analytical document.
(4) Data cube operations. This cube data generation step calculates data values
according to the dimension of the fact table for the subject-oriented analysis
requested by the query. The measurement values are first aggregated from the
XML-based cube data, which belongs to the same analytic dimension and then
used to calculate/obtain the results of subject analysis according to the specific
requirement of the query.
(5) Wrapping and delivery. Since XSLT is essential for converting online XML
documents to other browsable data formats (W3C, 2007), the last step wraps the
generated XML-based DC into a browsable XML-based DC by applying XSLT
and delivers it to users after authorisation and security checks.

Distributed decision support with authorisation and protection

For personalised decision support in a distributed web-based DWS, security
management is necessary, since decision resources required for business intelligence
are often sensitive for the enterprise. It is therefore necessary to balance the trade-off
between flexibility and authority in a web-based DWS. However we also need to
consider the trade-off between efficiency and security. Extending from the deployment of
a traditional database system adoption, this study used a push-and-pull mechanism to
perform authority and security management in the users’ support service component of
the web-based DWS. With a combination of user-pull and technical-push approaches,
users are presented only with information directly related to their corresponding
functions or interests. The option to receive additional information on demand is still
 OIR available. In addition users can still apply traditional OLAP tools to browse for data
available in DWS. Such architecture can therefore prioritise the corresponding security
36,1 requirements for various categories of users. Authorisation in the user support service
component is intended to allow users to access resources necessary for performing or
supporting their decision-making functions while restricting unauthorised access that is
not relevant to their decision-making needs (Stamper and Case, 2008). In addition to the
60 authorisation provided by the DWS service, a push-and-pull mechanism is a feasible
means of providing access restriction according to security policy in the user support
component. The levels of security requirements are classified according to the medium
used and also the security mechanism implemented.
.
Private. Within the most sensitive push-and-pull operations, email is used to
push decision information. To maintain privacy users’ requests should be pulled
with both email and DWS authentication, since email service is regarded as
Downloaded by JAZAN UNIVERSITY At 01:50 15 August 2017 (PT)

having a high degree of privacy and personalisation.

.
Protected. The protected push-and-pull operations provide medium security and
also two levels of authentication. Besides DWS authentication user-pull
operations can be authenticated by a web-based DWS front-end.
Decision-making information should be pushed to, or pulled by, users
according to web-based protocols for protection.
.
Public. For public or non-sensitive decisional tasks it is acceptable for users to
pull the requests directly by an OLAP application client. The only authentication
facilities are provided by a DWS. Decision information can therefore be pushed
to users by an OLAP application client.
User support service for ad hoc information. The user-pull operation is defined as
consisting of manual or ad hoc decisional or analytical requests from various types of
decision users. For ad hoc information on the one hand, the user support service will
identify data sources and authenticate the use of the resource according to security
levels. On the other hand the presentation transformation service retrieves data content
and schema from the data reservoir and converts it into an XML format before
presenting it to the users on demand.
User support service for routine information. The decision policy is pre-defined for
routine decisional or analytical requests from various types of decision users and tasks.
For routine information the user support service will again identify data sources and
authenticate the use of the resource according to security levels. In contrast the
presentation transformation service only retrieves data content from the data reservoir
and converts it with a pre-defined XML-based schema into an XML format before
presenting it to the users routinely. In summary the comparison of the unique
characteristics of on-demand information support and routine information support
operations is shown in Table I.

Assessment and validation

Flexibility and scalability are the underlying technological issues that must be considered
for the success or failure of a data warehouse. The web-based data warehouse is
increasingly under more stress than traditional data warehouses as the number of
decision users and associated tasks increases (for scalability) and the level of
sophistication of the browsing platforms rises (for flexibility) (Moeller, 2008). Security
 Support for
On-demand information support Routine information support
Decisional tasks Mainly for strategic tasks Mainly for operational tasks business
Frequency Occasionally Periodically
intelligence
Resource request Ad hoc Routine
Authentication and security Role dynamically defined Pre-defined application 61
policy application
Cube data transformation Transformation for both schema Transformation for content only
Table I.
strategy and content information according according to pre-defined schema
Characteristics of
to users’ ad hoc requirements template
information support
Metadata storage Run-time generation according to Pre-defined metadata stored in using the push-and-pull
XML-based schema on demand metadata repository mechanism
Downloaded by JAZAN UNIVERSITY At 01:50 15 August 2017 (PT)

may be another critical area for a data warehouse in particular (Moeller, 2008; Trujillo
et al., 2009a, b; Ahmad and Ahmad, 2010). The aim of a web-based data warehouse is to
provide an interactive analysis of warehouse data. Some of these data are proprietary and
highly sensitive by nature and consequently data security issues may be triggered during
the processes of sharing information and collaboration among users (Moeller, 2008).
This section concerns the validity of the proposed study with respect to “design
science” as discussed in the information systems literature (Hevner et al., 2004). For the
purpose of design evaluation, one can utilise system development, simulation, and
factor measurement to demonstrate the validity of the artefact. As discussed above this
study utilised the “design science” framework to provide a series of tests to establish
the feasibility of the proposed web-based DWS with an embedded push-and-pull
mechanism that facilitates the requirements of scalability, efficiency, security, and
personalisation in addition to needed flexibility. We provide a mathematical model to
assess the efficiency and scalability of the push-and-pull mechanism in web-based
DWS. Furthermore a simulation with an actual data set is used to assess the security
and personalisation of the proposed artefact.

Prototype development
To conduct a feasibility analysis of flexible business intelligence on the internet, we
developed a prototype web-based DWS derived and expanded from a previous study
with an embedded push-and-pull mechanism (Huang and Su, 2002). Extrapolating
from the previous study we have performed iterative tests to refine and extend
functionalities, including personalised user support services, multi-protocol delivery
and design facilities, and an event-based push-and-pull mechanism. Figure 4 shows the
back-end user-predefined cube schema designer component for the creation of an XML
cube metadata repository.
For routine decisional tasks to be triggered at the back-end, it is feasible to invoke a
pre-defined cube schema. This schema would be stored in an XML metadata repository
for routine utilisation by the presentation transformation service component. As a
flexible front-end it is feasible to use various media and devices, such as web browsers
or mobile devices, with applicable XML cube data while push-and-pull approaches are
established and authorisation is performed in the user support service component of
the proposed prototype system.
 OIR
36,1

62
Downloaded by JAZAN UNIVERSITY At 01:50 15 August 2017 (PT)

Figure 4.
Back-end cube schema
designer for user-pull
requirements
 Evaluation of efficiency and scalability for the push-and-pull mechanism Support for
In this section we will demonstrate the efficiency and scalability that can be achieved business
by calculating and comparing the corresponding overall costs between the traditional
on-demand mechanism and the push-and-pull mechanism on web-based DWS. The intelligence
traditional on-demand mechanism uses the same approach to retrieve a browsable
cube for various routine or ad hoc decisional requests; the push-and-pull mechanism
however decides to use the push or pull approach to retrieve a browsable DC according 63
to the decision types of such requests. To illustrate the distribution of decision types
involved, notice that the probabilities are 1 for the traditional decision mechanism,
f # 1 for the push mechanism, and 1-f for the pull mechanism. Value f is identified as
the ratio of request times of routine decisions to all decisions, which include both
routine and ad hoc decision types.
In the on-demand mechanism, as described in the section on the user support service
for ad hoc information, the overall costs of browsable cube retrieval on web-based
Downloaded by JAZAN UNIVERSITY At 01:50 15 August 2017 (PT)

DWS for decision task T with data dimension Q and result cube data set D can be
expressed as equation (1). NETWORK_PULL(T), ENCAPSULATION(T,D), and
NETWORK_PUSH(T,D) are the costs for delivering ad hoc requests from decision
users (user-pull operation), wrapping the generated information into a browsable DC
and returning it to users (technical-push operation). Authenticating the usage of the
resource by the user support service component is necessary for restricting
unauthorised access costs AUTH(T). Before an XML-based DC transformation, the
presentation transformation service costs DWS_QUERY(T,Q) to query DC are
identified according to the requirements of the requests from the back-end DWS. The
costs which the presentation transformation service employed for converting data
content and schema to XML format are DATA_TRANSFORMATION(T,D) and
SCHEMA_TRANSFORMATION(T,Q) respectively:

OVERALL_COST_FOR_ON _DEMANDðTÞ ¼ NETWORK_PULLðTÞ

þ AUTH ðTÞ þ DWS_QUERY ðT; QÞ þ SCHEMA_TRANSFORMATION ðT; QÞ
ð1Þ
þ DATA_TRANSFORMATION ðT; DÞ þ ENCAPSULATION ðT; DÞ
þ NETWORK_PUSH ðT; DÞ

In contrast with the embedded push-and-pull mechanism, as described in the sections

on user support services for ad hoc and routine information, the overall costs of
browsable cube retrieval on web-based DWS for decision task T with data dimension
Q and result cube data set D can be expressed as equation (2). The overall costs include
the costs of responding to both routine and ad hoc decision requests by taking
advantage of the push-and-pull approaches respectively. Since the decision policy is
pre-defined in the routine situation, the delivery of the routine decisional request will be
completed by an internal trigger of the mechanism costs INTERNAL_TRIGGER(T,D)
instead of NETWORK_PULL(T). Furthermore the cost of transforming cube schemas
into an XML format, SCHEMA_TRANSFORMATION(T,Q), is eliminated because an
XML-based schema is pre-generated for nonvolatile or routine DC. In the ad hoc
situation the embedded push-and-pull mechanism retrieves the browsable cube as in
the case of the on-demand mechanism, hence the costs are calculated in the same way
that the on-demand mechanism calculates them:
 OIR OVERALL_COST_FOR_PUSH _PULLðTÞ ¼
36,1 f £ ðINTERNAL_TRIGGERðT; DÞ þ AUTH ðTÞ þ DWS_QUERY ðT; QÞ
þ DATA_TRANSFORMATION ðT; DÞ þ ENCAPSULATION ðT; DÞ
þ NETWORK_PUSH ðT; DÞÞ þ ð1 2 f Þ ð2Þ
64 £ ðOVERALL_COST_FOR_ON _DEMANDðTÞÞ ¼ f
£ ðOVERALL_COST_FOR_PUSH _ONLY ðTÞÞ þ ð1 2 f Þ
£ ðOVERALL_COST_FOR_ON _DEMANDðTÞÞ

For efficiency and scalability of the push-and-pull mechanism, we can compare their
costs:
!
Downloaded by JAZAN UNIVERSITY At 01:50 15 August 2017 (PT)

Xn
lim
n!1
OVERALL_COST_FOR_ON _DEMANDðT K Þ =
k¼1
!
X
n
OVERALL_COST_FOR_PUSH _PULLðT k Þ ¼
k¼1
!
X
n
lim OVERALL_COST_FOR_ON _DEMANDðT K Þ = ð3Þ
n!1
k¼1
X
n
ð f £ OVERALL_COST_FOR_PUSH _ONLY ðT K Þ
k¼1


þð1 2 f Þ £ OVERALL_COST_FOR_ON _DEMANDðT K ÞÞ

where the number of tasks can run from T1 to Tn, and n can approach infinity.
If the value of equation (3) is larger than 1, the overall cost of the on-demand
mechanism is larger than the overall cost of the push-and-pull mechanism. This means
that efficiency is expected to be better with a push-and-pull mechanism in the
web-based DWS than with only an on-demand mechanism. This is mainly because it is
not necessary for pull operations to be performed by users manually and hence, the ad
hoc cube schema transformation can be done by the presentation transformation
service component to deal with routine decision tasks. The only additional costs of the
push-and-pull mechanism are the internal trigger costs for the cube generation
processes. The overall costs for the utilisation of the push-and-pull mechanism will be
lower, since for decision task T with data dimension Q and result cube data set D:

INTERNAL_TRIGGERðT; DÞ , NETWORK_PULLðTÞ
ð4Þ
þ SCHEMA_TRANSFORMATION ðT; QÞ

It should be further noted that a push-and-pull mechanism compared to a traditional

on-demand only operation will be more incrementally scalable and efficient as n ! 1
and f ! 1. Therefore the overall cost of the push-and-pull mechanism is less than the
overall cost of the traditional on-demand mechanism. This means that the
 push-and-pull mechanism can complete infinite decision tasks more efficiently and can Support for
facilitate scalability for the number of tasks flexibly.
business
Security and personalisation experience from a template data set intelligence
To assess the operational security and personalisation features of the proposed
architecture, we developed a simulation experiment by employing actual DC with the
underlying sample data of the annual sale and inventory history of an apparel mart in 65
2007. The following two factors, Degree_of_Security and Degree_of_Personalisation,
are defined to calculate the level of security and personalisation for the used external
schema under the push-and-pull mechanism prior to the on-demand approach for the
web-based DWS.
The sales DC is applied in the sample data set for the decision associated with
sales-related operations in 2007. The sales DC with fact table SALES_FACT_2007 stores
the facts measuring all details about sales and it contains foreign keys that refer to
Downloaded by JAZAN UNIVERSITY At 01:50 15 August 2017 (PT)

primary keys in its corresponding dimension tables as many-to-one relationships. There

are six dimension tables – Store, Customer, Promotion, Time_By_Day, and Product in
level 1 and Product_Class in level 2 – with each storing the attributes corresponding to
the specific aspects of a dimension. The inventory cube is also applied in the sample data
set for the decision regarding inventory-related operations in 2007. The inventory DC
with fact table INVENTORY_FACT_2007 stores the facts measuring all the details
about stock refreshed daily. The many-to-one relationships are defined between fact
table INVENTORY_FACT_2007 and its two-level corresponding dimension tables
including Store, Warehouse, Time_By_Day, Product, and Product_Class, with each table
storing attributes corresponding to the specific aspects of a dimension.
Previously we measured the degree of security as the ratio of data size for external
schema over data size for global schema, which identifies the portion of data size
authorised for specific decision makers D or tasks T:
Data_Size_for_Cube_SchemaðT; DÞ
Degree_of _SecurityðTÞ ¼ ð5Þ
Data_Size_for_Global_Schema
In this case the total data size of all tables actually accessible is calculated as:
2:3MB þ 1:8MB þ 146KB þ 133KB þ 33KB þ 9KB þ 4KB þ 115KB
þ 3KB ø 4:5MB
For decision makers or tasks referring to the sales cube, the values of data size for the
sales cube are illustrated in Table II and the security level of the assessable data
resource, the Degree_of_Security(SALES), is calculated as:
ð2:3MB £ 50% þ 1:8MB £ 20% þ 146KB £ 63% þ 133KB £ 46% þ 33KB £ 11%
þ 9KB £ 4% þ 4KB £ 28%Þ=4:5MB ø 37%
Similarly the values of data size for the inventory cube are illustrated in Table II and
the Degree_of_Security(INVENTORY) can be calculated as:

ð133KB £ 46% þ 33KB £ 11% þ 9KB £ 4% þ 4KB £ 28% þ 115KB * 62%

þ 3KB * 55%Þ=4:5MB ø 3%
 OIR
Cube schema
36,1 Table Data size Schema size (attribute length) Sales Inventory

Sales_Fact_2007 2.3 MB £ 50% 14 U

Customer 1.8 MB £ 20% 37 U
Promotion 146 KB £ 63% 50 U
66 Product 133 MB £ 46% 40 U U
Time_by_Day 33 KB £ 11% 5 U U
Product_Class 9 KB £ 4% 3 U U
Store 4 KB £ 28% 51 U U
Inventory_Fact_2007 115 KB £ 62% 18 U
Warehouse 3 KB £ 55% 66 U
Table II.
Data size and schema size Notes: Data size ¼ Estimated storage size of table £ estimate % of data length per record required by
for the cube schema the user or task
Downloaded by JAZAN UNIVERSITY At 01:50 15 August 2017 (PT)

Therefore 63 percent and 97 percent of the total data volume, respectively, are hidden
and inaccessible as the push-and-pull mechanism is performed for delivery of the sales
cube and the inventory cube.
The degree of personalisation can be measured as the ratio of an attribute length in
an external schema that belongs only to decision makers D or tasks T over the
attribute length in an external schema that belongs to all decision makers or tasks.

Schema_Size_only_for_Cube_SchemaðT; DÞ
Degree_of _PersonalzationðTÞ ¼ ð6Þ
Schema_Size_for_all_Cube_Schema
In this case the schema size used by any cube schema is calculated as:

14 þ 37 þ 50 þ 40 þ 5 þ 3 þ 51 þ 18 þ 66 ¼ 284ðbytesÞ
For decision makers or tasks referring to the sales cube, the values of schema size for
only the sales cube are illustrated in Table II and the personalisation level of assessable
data resource, Degree_of_Personalisation (SALES), is calculated as
14 þ 37 þ 50/284 ø 35.6%. The values of schema size only for the inventory cube
are illustrated in Table II and Degree_of_Personalisation (INVENTORY) can be
calculated as 18 þ 66=284 ø 29:6%. The above simulation for both decision tasks
shows that while 65.2 percent (35.6 percent þ 29.6 percent) of schema are personalised
and belong to individual decision tasks or users, only 34.8 percent (100 percent-65.2
percent) of schema will be shared among decision tasks.

Comparison with related studies of XML-based warehouses

Some typical approaches were proposed in related research for enabling and integrating
XML query languages with a native storage of the warehouse, such as Xyleme, XCube,
and DAWAX (Xyleme, 2001; Hümmer et al., 2003; Baril and Bellahsène, 2003). However
these approaches cannot be considered complete integration architectures for XML-based
warehouses. This is simply because these approaches only focus on integrating XML data
into web warehouses, modelling an XML data warehouse, and/or storing XML
documents in the data warehouses in a separate manner. Compared to our proposed
system Xyleme was aimed at gathering XML web sources and integrating them into a
 data warehouse without considering the associated characteristics of DC, while our Support for
mechanism allows users to store nonvolatile XML-based DC in the repository for efficient business
handling of routine operations. XCube was proposed to model an XML data warehouse
which stores facts and all dimensions in two respective XML documents. However not all intelligence
dimension information represented in the same XML document is efficient and flexible
(Boukraâ et al., 2006). Instead our mechanism stores facts and each dimension in one
single XML document respectively (see Figure 3) and by doing so, each dimension in one 67
XML document can be used to allow the best performance. This is because the
information on each dimension and its hierarchical levels can be updated more efficiently
and several facts can indeed share the same dimensions in a more flexible manner.
DAWAX was developed to consider only filtering and adapting gathered XML
documents according to user requirements by using the view mechanism. Consequently
it cannot be regarded as a comprehensive, personalised, decision supporting
Downloaded by JAZAN UNIVERSITY At 01:50 15 August 2017 (PT)

functionality by nature. Compared with DAWAX our mechanism provides proper

browsable cube retrieval approaches according to the requests generated by various
types of decision users and tasks by pursuing user-push and technical-pull strategies.
The increasing personalised decision tasks can therefore be completed in a more efficient
and scalable way. In addition XML-based warehouses are mainly used to facilitate the
exchange of a wide variety of data on the web in which security is one significant issue
(Moeller, 2008; Ahmad and Ahmad, 2010; Ravat et al., 2010). Our push-and-pull
mechanism also provides three authority and security management levels contrasting
with these aforementioned approaches. Personalised decision-making information can
therefore be properly modelled and securely delivered in different scenarios.

Case study
To ensure the practicality of our push-pull mechanism, this study applied it to a real
business case. The case company K is a wholesaler and retailer of clothing in Taiwan.
Company K offers more than 300 types of commodities and has 13 marketing branches
located in department stores. Having the most precise and latest sales and inventory
data in each marketing branch is critical to control real inventory stocks and support
decisions about sales plans. Each marketing branch has an anniversary sale every year
before Christmas to promote products. At this time of year the manager spends
considerable time making frequent round trips to and from marketing branches to
understand the overall sales situation. Retail staff cannot immediately gather
inventory information in each marketing branch when they want to allocate and
transfer goods among different marketing branches.
In order to take advantage of each selling opportunity and improve the effectiveness
of transferring goods among marketing branches, company K decided to analyse and
access the sales and inventory data among different marketing branches by using our
proposed mechanism system. Routine inventory information can be pushed to retail
staff periodically and efficiently to support operational tasks such as allocating goods
and transferring among marketing branches. In addition the manager who needs to
change the requirements of dimensional analysis of sales data can pull ad hoc requests
and the resulting mechanism system can push overall on-demand sales information to
support various types of strategic tasks such as the extension of the anniversary sale
deadline. Retail staff are not authorised to browse sensitive information. For instance
only the manager has the authority to browse the total sales figures for each marketing
 OIR branch. Generated sales or inventory information are expressed in XML documents
36,1 and wrapped into the browsable formats by applying XSLT and then this information
is freely propagated to controlled mobile devices provided by company K. By doing so
both the manager and retail staff can therefore get information efficiently, understand
the latest sales and inventory situations in a timely manner, and make the right
decisions in real time. Figure 5 shows the sales information presented with open XML
68 expressions as viewed on a mobile device by the manager.

Conclusions
This research provided an architectural solution for web-based DWS construction with
XML structural metadata to overcome the limitations of current systems for supporting
a personalised distributed decision making/supporting mechanism. It provides two
additional service modules – the user support service and the presentation
Downloaded by JAZAN UNIVERSITY At 01:50 15 August 2017 (PT)

transformation service – in order to balance such data requirements as heterogeneity,

shareability, security, and privacy. In such architecture dimensional information will be
open and public XML expressions as a possible solution to the complexity problem of
heterogeneous data integration among DWS and other data resources. Furthermore with
the push-and-pull operations as a fundamental element to support personalised
decision-making activities, users are able to be informed either on demand or with only
the information directly related to corresponding decisional scenarios.
To efficiently, flexibly, and rapidly apply business intelligence to key business
problems, it is necessary to improve the quality of online information and the speed of
decision-making in the e-business environment. The quality of online information
refers to the believability of information and/or its source (Metzger, 2007). However it is

Figure 5.
Browsable sales cube data
pushed to mobile devices
 hampered by a contemporary business environment in, which increasingly copious Support for
volumes of distributed and heterogeneous data and various decision makers/tasks business
have to interact. To overcome these problems and achieve their business objectives,
more and more companies have begun to take advantage of technology to cope with intelligence
the multitude of decisional resources and tasks, and acquire the potential knowledge to
fight off challenges. In addition the requirements of decision tasks are not always
stable in this dynamic and fast changing world. Believable online information – 69
knowledge provided for various decision makers to solve different business problems
– must be syntactic for structural consistency, semantic for proper meaning, and
pragmatic for appropriate purposes (Will, 2006). It remains very difficult in the current
internet-based DWS environment to support various decision makers in efficient ways
with appropriate and understandable information.
Therefore the proposed architecture embedded a push-and-pull mechanism to
Downloaded by JAZAN UNIVERSITY At 01:50 15 August 2017 (PT)

support users in various situations and meet these three dimensions of online
information quality. In routine situations on the one hand the mechanism can push the
right information efficiently to appropriate decision makers to protect sensitive data
according to pre-defined authorisation and security requirements as specified in a meta
schema. On the other hand decision makers with changing requirements can also pull
ad hoc requests and the mechanism can push on-demand information. The user-pull
and technical-push processes in ad hoc situations are also authorised according to the
organisational role of decision makers. Therefore, by taking advantage of the
mechanism, various decision makers can efficiently get the proper information
syntactically, understand meanings of information semantically, and fulfil their
purpose by using this information pragmatically. Consequently the proposed
architecture provides a support mechanism for business intelligence to efficiently
and flexibly help companies make the right decisions in real time, grasp business
opportunities, and gain competitive advantage.
As for future research on distributed decision support mechanisms, we suggest
focusing on decision-supporting mechanisms in various types of distributed system
topologies. For example peer-to-peer and grid topologies might be utilised in virtual
organisations or communities. The study makes assumptions that the
data/information processed in the mechanism are not ambiguous, although proper
applications of the mechanism can help to eliminate syntactic and semantic
ambiguities. With both rigorous schema definitions/expressions and explicit user
requests, the mechanism can therefore provide decisional information more precisely
and in a more timely way. However decision makers must still look for the best
solution to their problems and fully understand the “pushed-up” answers in the ad hoc
situation in response to the “pull-up” questions. Likewise decision supporters must
fully understand the decision situation when designing routine solutions (Will, 2006).

References
Ahmad, S. and Ahmad, R. (2010), “An improved security framework for data warehouse: a hybrid
approach”, Proceedings of the International Symposium on Information Technology in
Kuala Lumpur, IEEE, Los Alamitos, CA, pp. 1586-90.
Ariyachandra, T. and Watson, H. (2010), “Key organizational factors in data warehouse
architecture selection”, Decision Support Systems, Vol. 49 No. 2, pp. 200-12.
 OIR Baril, X. and Bellahsène, Z. (2003), “Designing and managing an XML warehouse”, in Chaudhri, A.B.,
Rashid, A. and Zicari, R. (Eds), XML Data Management: Native XML and XML-enabled
36,1 Database Systems, Addison Wesley, New York, NY, pp. 455-73.
Bhargava, H.K., Power, D.J. and Sun, D. (2007), “Progress in web-based decision support
technologies”, Decision Support Systems, Vol. 43 No. 4, pp. 1083-95.
Boukraâ, D., Messaoud, R.B. and Boussaid, O. (2006), “Proposition d’un modèle physique pour les
70 entrepôts XML”, Proceedings of Atelier Systèmes Décisionnels en conjonction avec
9th Maghrebian Conference on Information Technologies in Agadir, Morocco, available at:
http://boukraa.freehostia.com/publications/ASD2006_Boukraa.pdf (accessed 23 October
2011).
Cheung, W. and Babin, G. (2006), “A metadatabase-enabled executive information system (Part B):
methods for dynamic multidimensional data analysis”, Decision Support Systems, Vol. 42
No. 3, pp. 1599-612.
Downloaded by JAZAN UNIVERSITY At 01:50 15 August 2017 (PT)

Chung, Y.D., Yang, W.S. and Kim, M.H. (2007), “An efficient, robust method for processing of
partial top-k/bottom-k queries using the RD-Tree in OLAP”, Decision Support Systems,
Vol. 43 No. 2, pp. 313-32.
Eavis, T. (2009), “Parallel and distributed data warehouses”, Encyclopedia of Database Systems,
Springer, New York, NY.
El-Gayar, O. and Tandekar, K. (2007), “An XML-based schema definition for model sharing and
reuse in a distributed environment”, Decision Support Systems, Vol. 43 No. 3, pp. 791-808.
Furtado, P. (2009), “A survey of parallel and distributed data warehouses”, International Journal
of Data Warehousing and Mining, Vol. 5 No. 2, pp. 57-77.
Grabova, O., Darmont, J., Chauchat, J.H. and Zolotaryova, I. (2010), “Business intelligence for
small and middle-sized enterprises”, SIGMOD Record, Vol. 39 No. 2, pp. 39-50.
Hevner, A.R., March, S.T., Park, J. and Ram, S. (2004), “Design science in information systems
research”, MIS Quarterly, Vol. 28 No. 1, pp. 75-105.
Huang, S.M. and Su, C.H. (2002), “The development of an XML-based data warehouse system”,
Proceedings of Intelligent Data Engineering and Automated Learning in Manchester,
Springer-Verlag, London, pp. 206-12.
Huang, S.M., Chou, T.H. and Seng, J.L. (2007), “Data warehouse enhancement: a semantic cube
model approach”, Information Sciences, Vol. 177 No. 11, pp. 2238-54.
Hümmer, W., Bauer, A., Harde, G. and XCube:, X.M.L. (2003), “XCube: XML for data
warehouses”, Proceedings of the 6th International Workshop on Data Warehousing and
OLAP in New Orleans, ACM Press, New York, NY, pp. 33-40.
Kumar, N., Gangopadhyay, A., Bapna, S., Karabatis, G. and Chen, Z. (2008), “Measuring
interestingness of discovered skewed patterns in data cubes”, Decision Support Systems,
Vol. 46 No. 1, pp. 429-39.
McFadzean, E., Ezingeard, J. and Birchall, D. (2007), “Perception of risk and the strategic impact
of existing IT on information security strategy at board level”, Online Information Review,
Vol. 31 No. 5, pp. 622-60.
Mahboubi, H., Ralaivao, J.C., Loudcher, S., Boussaid, O., Bentayeb, F. and Darmont, J. (2009),
“X-WACoDa: an XML-based approach for warehousing and analyzing complex data”,
in Bellatreche, L. (Ed.), Data Warehousing Design and Advanced Engineering Applications:
Methods for Complex Construction, Idea Group Publishing, New York, NY, pp. 38-54.
Mehedinţu, A., Buligiu, I. and Pı̂rvu, C. (2008), “Web-enabled data warehouse and data
webhouse”, Revista Informatica Economic Journal, Vol. 1 No. 45, pp. 96-102.
 Metzger, M.J. (2007), “Making sense of credibility on the web: models for evaluating online Support for
information and recommendations for future research”, Journal of the American Society for
Information Science and Technology, Vol. 58 No. 13, pp. 2078-91. business
Moeller, R.A. (2008), Distributed Data Warehousing Using Web Technology: How to Build a More intelligence
Cost-effective and Flexible Warehouse, AMACOM, New York, NY.
Møller, A. and Schwartzbach, M.I. (2006), An Introduction to XML and Web Technologies,
Addison-Wesley, Harlow. 71
Ravat, F., Teste, O., Tournier, R. and Zurfluh, G. (2010), “Finding an application-appropriate
model for XML data warehouses”, Information Systems, Vol. 35 No. 6, pp. 662-87.
Stamper, D.A. and Case, T.L. (2008), Business Data Communications, 6th ed., Prentice Hall,
Upper Saddle River, NJ.
Trujillo, J., Soler, E., Fernández-Medina, E. and Piattini, M. (2009a), “A UML 2.0 profile to define
security requirements for data warehouses”, Computer Standards & Interfaces, Vol. 31
No. 5, pp. 969-83.
Downloaded by JAZAN UNIVERSITY At 01:50 15 August 2017 (PT)

Trujillo, J., Soler, E., Fernández-Medina, E. and Piattini, M. (2009b), “An engineering process for
developing secure data warehouses”, Information and Software Technology, Vol. 51 No. 6,
pp. 1033-51.
Tummarello, G., Morbidoni, C., Puliti, P. and Piazza, F. (2008), “A proposal for textual encoding
based on semantic web tools”, Online Information Review, Vol. 32 No. 4, pp. 467-77.
Vine, D. (2000), Internet Business Intelligence: How to Build a Big Company System on a Small
Company Budget, Cyberage Books, Medford, NJ.
W3C (2007), “XSL Transformations (XSLT) version 2.0”, available at: www.w3.org/TR/xslt20/
(accessed 23 October 2011).
Will, H. (2006), “Knowledge management and administration depend on semiotic information
systems”, International Journal of Management and Decision Making, Vol. 7 No. 1, pp. 36-57.
Wu, D., Cheong, C.H. and Wong, M.H. (2009), “Supporting asynchronous update for distributed
data cubes”, Journal of Network and Computer Applications, Vol. 32 No. 4, pp. 889-900.
Xyleme, L. (2001), “A dynamic warehouse for XML data of the web”, IEEE Data Engineering
Bulletin, Vol. 24 No. 2, pp. 40-7.

About the authors

Jing-Shiuan Hua is a PhD student in the Department of Information Management at National
Chung Cheng University in Taiwan. Her current research interests include software engineering,
project management and information security management.
Shi-Ming Huang is Head of the Accounting and Information Technology Department and a
Director of the Research Centre of e-Manufacturing and e-Commerce at National Chung Cheng
University. He was the 2006 President of the International Chinese Information Systems
Association. He has published five books, three business software products, and over 40 refereed
articles. He has served as an editorial board member for several international journals and has
acted as a consultant for a variety of Taiwanese government departments, software companies
and other companies.
David C. Yen is a Raymond E. Glos Professor of Business and a Professor of MIS in the
Department of Decision Sciences and Management Information Systems at Miami University. He
is active in research and has published books and articles. His research interests include data
communications, electronic/mobile commerce, databases, and systems analysis and design.
David C. Yen is the corresponding author and can be contacted at: [email protected]

To purchase reprints of this article please e-mail: [email protected]

Or visit our web site for further details: www.emeraldinsight.com/reprints
 This article has been cited by:

1. Saeed Rouhani Faculty of Management, University of Tehran, Tehran, Iran Amir Ashrafi Department
of Industrial Management, Allameh Tabataba’i University, Tehran, Iran Ahad Zare Ravasan Department
of Industrial Management, Allameh Tabataba’i University, Tehran, Iran Samira Afshari Department
of Industrial Management, Allameh Tabataba’i University, Tehran, Iran . 2016. The impact model of
business intelligence on decision support and organizational benefits. Journal of Enterprise Information
Management 29:1, 19-50. [Abstract] [Full Text] [PDF]
2. Sean M. Stone, M. Sara Lowe. 2014. Who is Citing Undergraduate Theses in Institutional Digital
Repositories? Implications for Scholarship and Information Literacy. College & Undergraduate Libraries
21:3-4, 345-359. [CrossRef]
3. Susan E. Hahn, Anna Wyatt. 2014. Business Faculty's Attitudes: Open Access, Disciplinary Repositories,
and Institutional Repositories. Journal of Business & Finance Librarianship 19:2, 93-113. [CrossRef]
4. Kijpokin KasemsapImplementing Business Intelligence in Contemporary Organizations 177-192.
Downloaded by JAZAN UNIVERSITY At 01:50 15 August 2017 (PT)

[CrossRef]
5. Kijpokin KasemsapThe Role of Data Mining for Business Intelligence in Knowledge Management 12-33.
[CrossRef]
6. Kijpokin KasemsapImplementing Business Intelligence in Contemporary Organizations 33-48. [CrossRef]