Mobile Networks and Applications 7, 185198, 2002

2002 Kluwer Academic Publishers. Manufactured in The Netherlands.

Mobile Commerce: Framework, Applications and Networking

Support
UPKAR VARSHNEY
Department of Computer Information Systems, Georgia State University, Atlanta, USA

RON VETTER
Department of Computer Science, University of North Carolina, Wilmington, USA

Abstract. Advances in e-commerce have resulted in significant progress towards strategies, requirements, and development of e-commerce
applications. However, nearly all e-commerce applications envisioned and developed so far assume fixed or stationary users with wired
infrastructure. We envision many new e-commerce applications that will be possible and significantly benefit from emerging wireless and
mobile networks. To allow designers, developers, and researchers to strategize and create mobile commerce applications, we propose a fourlevel integrated framework for mobile commerce. Since there are potentially an unlimited number of mobile commerce applications, we
attempt to identify several important classes of applications such as mobile financial applications, mobile inventory management, proactive
service management, product location and search, and wireless re-engineering. We discuss how to successfully define, architect, and
implement the necessary hardware/software infrastructure in support of mobile commerce. Also, to make mobile commerce applications a
reality, we address networking requirements, discuss support from wireless carriers, and present some open research problems.
Keywords: mobile commerce, layered framework, mobile applications, middleware, wireless networking

1. Background and motivation

Electronic commerce has attracted significant attention in the
last few years. This high profile attention has resulted in significant progress towards strategies, requirements, and development of e-commerce applications [8]. The growth forecast for both business-to-consumer (B2C) and business-tobusiness (B2B) aspects of e-commerce over the next few
years is phenomenal by any standard. One point that should
be made here is that nearly all e-commerce applications envisioned and developed so far assume fixed or stationary users
with wired infrastructure, such as a browser on a PC connected to the Internet using phone lines or a Local Area Network (LAN). We envision many new e-commerce applications will be possible and significantly benefit from emerging
wireless and mobile networks. We term these applications
wireless e-commerce or mobile commerce.
Wireless and mobile networks have experienced exponential growth in terms of capabilities of mobile devices, middleware development, standards and network implementation,
and user acceptance [18]. Currently, more than 800 million
cell phones and other mobile devices are in use worldwide,
and out of those, more than 140 million users are in US alone
(www.wow-com.com). The worldwide numbers are projected to rise to 1 billion soon, thereby exceeding the combined total of all computing devices several fold. In addition, countries with a lack of regular telecom infrastructure
are likely to adopt wireless and mobile communications to
serve both urban and rural areas. According to estimates by
GartnerGroup, in 2004, at least 40% of business-to-consumer
e-commerce will be initiated from smart phones supported by

WAP (Wireless Application Protocol). A study from the Wireless Data and Computing Service, a division of Strategy Analytics, reports that the mobile commerce market may rise to
$200 billion by 2004. The report predicts that transactions
via wireless devices will generate about $14 billion a year.
In this article, we examine how new m-commerce applications can be designed and supported by wireless and mobile
networks and mobile middleware. How well these applications become adopted by a business will depend on how fast
these applications can be deployed, the costvalue ratio, acceptance of new technologies by users and businesses based
on easy to use and uniform interfaces, and the building of trust
necessary to conduct m-commerce transactions while on the
move. We strongly believe that with the widespread deployment of wireless technologies, the next phase of electronic
business growth will be in the area of wireless and mobile
e-commerce. We are aware that consensus within business
and industry of such future applications is still in its infancy.
However, we are interested in examining those future applications and technologies that will form the next frontier of
electronic commerce.
To help future applications and technologies handle
m-commerce, we propose the framework as shown in figure 1. This framework will allow developers and providers
to strategize and effectively implement mobile commerce applications. The framework defines multiple functional layers,
simplifying the design and development, so different parties
(vendors, providers, and designers, etc.) can focus on individual layers. By following this framework, a single entity is not
forced to do everything to build m-commerce systems, rather,
they can build on the functionalities provided by others. This

186

VARSHNEY AND VETTER

Figure 2. Mobile commerce life cycle.

Figure 1. A framework for m-commerce.

will speed up the development of m-commerce applications

as designers and developers can assume that certain functions
will be provided by lower layers, and therefore, they need not
focus on the capabilities and constraints of individual devices
and networks.
This framework has four levels: m-commerce applications, user infrastructure, middleware, and network infrastructure. The framework shows that the design of new
mobile commerce applications should take into consideration the general capabilities of user infrastructure (mobile devices), and not the individual devices. With its ability to hide
details of underlying wireless and mobile networks from applications while at the same time providing a uniform and
easy to use interface, mobile middleware clearly is an extremely important component in developing new mobile commerce applications. The network infrastructure also plays an
important role in mobile commerce, as the user perceived service quality depends on available resources and capabilities
of wireless and mobile networks.
An open framework will prevent the design and development of proprietary products and services that may be built
in an ad hoc fashion. We believe that our framework will
allow interoperability of m-commerce applications and products from different providers and vendors. This would help in
the adoption of m-commerce on a global scale.
The framework also provides a developer and provider
plane to address the different needs and roles of application
developers, content providers, and service providers. Each
one of these could build its products and services using the
functionalities provided by others. A content provider can
build its service using applications from multiple application
developers. They can also aggregate content from other content providers and can supply the aggregated content to a network operator or service provider. Service providers can also
act as content aggregators, but are unlikely to act as either
an application or content provider due to their focus on the
networking and service aspects of m-commerce. A service
provider can also act as a clearing house for content and application providers in advertising and distributing their products
to its customers. In any case, the developer and provider plane
in our framework is likely to have multiple layers. Wireless
carriers can play a very active and important role in the mo-

bile commerce applications and services due to the fact that

a mobile user is going through their networks to perform all
mobile commerce transactions. Additionally, a mobile user
is likely to prefer one common bill (bundled services) for
voice, data, and mobile commerce services. However, there
are many technical and non-technical hurdles (such as pricing for mobile commerce transactions) that need to be overcome before carriers can become major players in this emerging field. Many of these players and possible interactions are
shown in figure 2.
We now turn to mobile commerce applications (section 2),
followed by what is needed to make it a reality sections
on user infrastructure and mobile middleware (section 3) and
wireless network infrastructure (section 4). Then we discuss
issues for wireless carriers and applications developers (section 5). We present several interesting research problems in
mobile commerce (section 6). Finally, in section 7, we make
some concluding remarks.
2. Emerging mobile commerce applications
Since there are potentially an unlimited number of mobile
commerce applications, we attempt to identify several important classes of applications and provide examples within each
class. Table 1 shows many such classes and example applications.
According to estimates from a report by Durlacher [9],
more than half of the European mobile commerce market in
the next few years may consist of financial, advertising and
shopping services. In this paper, however, we attempt to cover
a comprehensive range of mobile commerce applications under different classes with varying requirements in terms of
devices, middleware, and networks.
In the following subsections, we discuss mobile financial
applications, mobile advertising, mobile inventory management, proactive service management and other application
classes in more detail. We provide some scenarios and include a discussion on the required infrastructure and related
business issues.
2.1. Mobile financial applications (MFA)
Mobile financial applications are likely to be one of the most
important components of m-commerce. These could involve

MOBILE COMMERCE

187

Table 1
Details and networking requirements of m-commerce applications.
Class of applications
Mobile financial applications (B2C, B2B)
Mobile advertising (B2C)
Mobile inventory management (B2C, B2B)/
Product locating and shopping (B2C, B2B)

Proactive service management (B2C, B2B)

Wireless re-engineering (B2C, B2B)

Mobile auction or reverse auction (B2C,

B2B)
Mobile entertainment services and games
(B2C)
Mobile office (B2C)
Mobile distance education (B2C)
Wireless data center (B2C, B2B)

Details

Examples

Applications where mobile device becomes a powerful financial medium.

Applications turning the wireless infrastructure and
devices into a powerful marketing medium.
Applications attempting to reduce the amount
of inventory needed by managing in-house and
inventory-on-move.
Applications helping to find the location of product
and services that are needed.
Applications attempting to provide users information on services they will need in very-near-future.
Applications that focus on improving the quality of
business services using mobile devices and wireless
infrastructure.
Applications allowing users to buy or sell certain
items using multicast support of wireless infrastructure.
Applications providing the entertainment services
to users on per event or subscription basis.
Applications providing the complete office environment to mobile users any where any time.
Applications extending distance/virtual education
support for mobile users everywhere.
Applications supporting large amount of stored data
to be made available to mobile users for making intelligent decisions.

Banking, brokerage, and payments for mobile

users.
User specific and location sensitive advertisements.
Location tracking of goods, boxes, troops, and
people.
Finding the location of a new/used car of certain
model, color and features.
Transmission of information related to aging (automobile) components to vendors.
Instant claim-payments by insurance companies.

Airlines competing to buy a landing time slot during runway congestion (a proposed solution to
air-traffic congestion problem).
Video-on-demand, audio-on-demand, and interactive games.
Working from traffic jams, airport, and conferences.
Taking a class using streaming audio and video.
Detailed information on one or more products
can be downloaded by vendors.

Note: B2C: business-to-consumer, B2B: business-to-business.

Figure 3. Several mobile financial services.

a variety of applications such as mobile banking and brokerage service, mobile money transfer, and mobile micropayments as shown in figure 3. These services could turn
a mobile device into a business tool, replacing bank, ATM,
and credit cards by letting a user conduct financial transactions with mobile money. Certainly, more work is needed in
providing transaction support in the applications and network
infrastructure. Secure transactions are required before any of
these applications are widely deployed.
One interesting mobile financial application is micropayment involving small purchases such as vending and other
items. A mobile device can communicate with a vending
machine using a local wireless network to purchase desired
items. Micro-payments can be implemented in a variety of
ways. One way is that the user could make a call to a certain
number where per-minute charges equal the cost of vending

item. This approach has been used by SONERA, a Finish

wireless provider, in their famous Coke (and now Pepsi) machine service [9]. In effect, it collects money from the users
and credits it to the vending providers. Another way to perform micro-payments may be via using pre-paid numbers purchased from a service provider, bank, or credit-card company.
To support financial transactions including micro-payments, a
mobile service provider could act as a bank, acquire a bank,
or compete with a bank. There are many questions that have
to be addressed. One such question is the real cost of mobile micro-payments. Another question is how would a phone
company or other payer make any profit on mobile micropayments? There are several possible answers to this question. First, service providers could require pre-payments by
users, thereby leading to some financial profits. Also, the
cost of micro-payments may decrease with increased numbers of transactions or users. A provider can also charge a
small amount for the payment service or can absorb the cost
in order to provide this micro-payment as a competitive feature leading to increased number of customers and possibly
higher revenues for wireless provider.
2.2. Mobile advertising
Mobile advertising is also a very important class of mcommerce applications. Using demographic information collected by wireless service providers and information on the
current location of mobile users, very targeted advertising can
be done. The advertising messages can be personalized based

188

VARSHNEY AND VETTER

Figure 4. A possible scenario for mobile advertising and shopping.

on information provided by consulting the user at an earlier

stage or by the history of users purchasing habits. Advertisements sent to a user can also be location-sensitive and can
inform a user about various on-going specials (shops, malls,
and restaurants) in surrounding areas as shown in figure 4.
This type of advertising can be performed using Short Messaging Service (SMS) or by using short paging messages to
mobile users. The messages can be sent to all users located in
a certain area (the geographic region can be identified by advertisers or even by users in advance), a user-specific message
can be sent independent of the users current location.
As more wireless bandwidth becomes available, content
rich advertising involving audio, pictures and video clips can
be produced for individual users with specific needs, interests,
and inclinations. Also depending on interests and the personality type of individual mobile users, a network provider may
consider using a push or pull method of mobile advertising based on a per-user basis or a class of users. Other interesting issues include the number of advertisements and the
level and type of content that should be transmitted. These
advertisements should be limited to avoid overwhelming the
user with information and also to avoid the possibility of congestion of the wireless links. Wireless networks may consider
such advertising lower priority traffic if network load crosses
a certain threshold. Since these services need the current location information of a user, a third party may be needed to
provide location services. However, this may require a sharing of revenues between the network service provider and location service provider. One very interesting question here
is who owns the location information and in what ways such
information can be shared between wireless service providers
and others. Other issues include the amount of storage (user
location and profiles) and real-time processing requirements
to capture user information in a high variability environment.
The dependability of infrastructure (database/networks) will
also play an important role. The impact of a message transmission to an unintended receiver should also be considered.
Many different ways to charge for advertising (size of mes-

sage, number of customers, value of the items to be sold) can

be used. It is also possible that direct advertising to users may
be performed without much control from the wireless service
providers.
2.3. Mobile inventory management (MIM)
This class of applications involves location tracking of goods,
services, and possibly even people. The tracking of goods
may help service providers in determining the time of delivery to customer, thus improving customer service and obtaining a competitive edge over other businesses. One very interesting application is rolling inventory which may involve
multiple trucks carrying a large amount of inventory while on
the move. Whenever a store needs certain items/goods, it can
locate a truck (preferably in nearby area) and just-in-time delivery of goods can be performed. The rolling inventory and
delivery application can reduce the amount of inventory space
and cost for both vendors and stores and may also reduce the
time between when an order is placed and the goods are delivered (figure 5). Rolling inventory is a B2B m-commerce
application while location tracking can be considered a B2C
application. Using inexpensive embedded radio/microwave
devices (chips), a wireless network can track goods and services. Since satellite signals may not work well inside a truck,
a separate wireless LAN can be provided on-board for intratruck communication and tracking. An interesting research
problem is to determine an appropriate match for the amount
of inventory carried by trucks in a geographical area with dynamically changing delivery demands. Also traffic in a city
may affect the just-in-time delivery in nearby areas.
Another example of MIM is just-in-time delivery/movement of components in an assembly plant based on the rate of
consumption of existing components (figure 6). A variety of
new components can be moved a certain speed after receiving
a wireless signal from the components reaching the assembly
line or from a device on the assembly line itself. This will
allow just-in-time delivery leading to a reduced inventory and

MOBILE COMMERCE

189

Figure 5. Location tracking of goods.

Figure 6. Location tracking and just-in-time movement of components.

assembling cost. If the new components are delayed for some

reason, then signals can be sent to the assembly line for possible adjustment of the assembly speed to match the arrival
time of new components. Such an application would reduce
the inventory cost while increasing productivity by matching
the speed of new component arrival to the rate of assembly.
For implementation purposes, location tracking of components can be broken into two components: indoor and outdoor. Indoor tracking can be performed by a chipset (Tx/Rx)
and location information may be transmitted over a satellite
or cellular/PCS system to the component supplier where such
information is needed.
Implementation considerations in MIM include cost, reliability of wireless infrastructure, and the level of comfort with
a new technology. Since these applications (especially the
rolling inventory application) may present a paradigm shift in
how inventories are managed today, an interesting research
problem would be a study involving the cost and availability comparison of two inventory systems (regular and rolling)
under dynamically changing demands. Potential MIM customers may include shipping companies, assembly plants
(auto, manufacturing), airline/mass-transit industry, and supermarket chain stores. One positive factor is that many of
these industries are already increasing their use of wireless
technologies.

Possible wireless technologies that can be used with MIM

include Global Positioning Satellite Systems (GPS) that are
operated by the Department of Defense for location tracking
(the accuracy level is based on the type of receiver and the
type of user). Many luxury cars already have built-in GPS receivers that compute location information based on received
signals from at least 4 (out of 24) GPS satellites. GPS may
not work well for indoor environments, as they require line of
sight transmission between devices and satellites. Another location tracking application is called E (enhanced) 911. E-911
allows operators to receive location information of cell phone
users along with their emergency calls [5]. Since it is difficult for cellular/PCS systems to provide such location information, it is likely that a GPS chip will be installed on cell
phones to help 911 operators to determine the location of mobile users. If such a chip is installed in cell/PCS phones, then
location tracking of such users under normal circumstances
will also be possible. Since, people can also be considered
inventory resources, they could also be better managed.
2.4. Product location and search (PLS)
This class of applications includes locating an item in a particular area or location. This is somewhat different from
the previous class of applications as here we are concerned
with finding an item (or person) with certain specifications,

190

VARSHNEY AND VETTER

Figure 8. Information transmission on aging automobile components to

dealers.
Figure 7. Product location and search.

and whether it is available in a specified area or not. Potentially, there could be multiple places where such an item or an
item of similar attributes (also specified by the user) are located. Currently, many people go to several stores to find an
item (certain brand/size of TV, VCR or an automobile) and
compare price and features. Using a mobile device (such as
PalmPilot, Nokia Communicator or Net Phone) and a centralized/distributed database containing information on products,
a user should be able to find the exact location of a store where
a certain item is located. A list of locations and distance from
a specified point can also be displayed. After that the user can
buy on-line using a browser on his/her mobile device. In the
case of multiple stores/vendors carrying an item desired by
a user, they could compete to get the customer by real-time
manipulation of prices or by offering instant discounts.
From a technological point of view, a mobile user can send
a query message to a centralized location (shown in figure 7),
which in turn can interface several different stores/dealers and
decide if the item is available or not (and if yes, at what
price). Alternatively, stores/vendors may connect their inventory record systems to this site. Since the inventory of different vendors may use different code names, a uniform product
naming system (or existing code such as UPC) that allows
for easy translation to standard web content will be required.
If a database is not employed, the mobile user may need to
query the stores directly. The amount of wireless traffic may
become a problem if the total number of queries per item per
user exceeds the capacity of the wireless infrastructure. To
avoid high traffic levels, one may prefer to ship only product
codes rather than entire data sets.
Two factors to consider are: (a) how the database will price
its services to vendors/dealers and (b) the correctness of information (related to availability or price) from inventory to the
database or website. We also believe that software agent technologies will prove to be invaluable as multiple agents can
be deployed (cooperating & negotiating) to conduct various
transactions at different places.
2.5. Proactive service management (PSM)
This class of applications is based on collecting pertinent information about current or near-future user needs and providing services to users proactively. One such application

may involve collecting information about the aging components of an automobile (shown in figure 8). Many vendors
including car dealers/repair shops can increase their business
by acquiring information about aging components of an automobile. Information can be collected and used by car dealers for ordering components (thus reducing inventory costs).
In a more elaborate scenario, several dealers/shops can compete for business by offering discounts or lower rates. Such
information can also be collected and analyzed by manufactures to improve the design and manufacturing of future
products. This service could even be offered as part of the
warranty for new cars or could be purchased by current automobile owners. This would help reduce anxiety levels of
owners and improve the general conditions of automobiles on
the road leading to a reduced number of traffic jams, accidents, and even fatalities. From a technological point of view,
automobiles can be equipped with smart sensors that keep
track of how much wear and tear a car component has gone
through. This information can then be transmitted using a
radio/microwave/satellite system to a specified service center
or other location. Some implications of such applications are
privacy, security, reliability, and cost of deployment. However, we envision that car dealers, repair shops, automobile
owners, and even law enforcement officers (from the public
safety point of view) would be interested in such an application.
2.6. Wireless business re-engineering
Many insurance business situations involve going to customers premises, taking notes of a particular situation, going back to the office, and then taking suitable actions. This
process takes a long time and is not very efficient. To improve
this business practice, a wireless business re-engineering application can be used which allows on-the-spot claim adjustment. In such a scenario, a claim adjuster goes to the customers place, takes pictures and stores them in his/her mobile device for uploading to the companys database. Next the
adjuster downloads necessary information (customers profile
and coverage information) from insurance companys database. Using a small printer attached to the mobile device,
he/she prints a claim check. This entire process is performed
in minutes as opposed to days. The speed of this kind of service can significantly add to a companys bottom line and
competitive advantage. Security and connectivity issues are

MOBILE COMMERCE

191

important as they may affect the perceived quality of service

by a customer.

multi network interfaces for increased and reliable wireless access,

2.7. Mobile auction, entertainment and other services

basic security features to handle malicious code, support

for authenticating user, servers, and applications,

With an increasingly mobile society, more and more people

are on the move. While mobile, people may prefer to be involved in some business or entertainment activities. Many
of these services can be offered to people through mobile
devices and wireless networks. These include mobile auction/reverse auction, video-on-demand services, and other
entertainment-oriented services. The technologies needed include mobile devices with capabilities to match desired applications, suitable mobile middleware, and wireless networks
with high bandwidth (such as emerging LEO satellites or third
generation wireless networks). Continued connectivity is a
real important issue as it may affect the perceived quality
of service for entertainment/information services. For auction/reverse auction, frequent disconnection may seriously affect the usefulness of this service unless it can be guaranteed that if users get disconnected, the state of auction will
be maintained and disconnected users will not suffer any loss
during periods of disconnection.
3. Wireless user infrastructure and middleware issues
To make the applications described previously a reality, several functional components are necessary as highlighted in
figure 1. One important area includes mobile devices with
sufficient power in terms of memory, display, and communications functionalities. The hand-held device is really the
entry point for most mobile commerce systems. The capabilities or limitations of these devices will impact the type
and frequency of which mobile commerce applications will
be used.
The devices available today [10] can be characterized as
either communication-centric or computing-centric, however,
in near future such differences may cease to exist as these devices converge to a single intelligent mobile device. Many
hand-held devices now support a variety of network interfaces and access ranges, from short-range Bluetooth, to wireless LAN, to wireless WAN access. Location support, video
streaming, barcode readers and other features are also being
implemented in many of the devices.
Although it is possible to do some simple mobile commerce transactions using a simple hand-held device, many
sophisticated mobile commerce applications require other capabilities. For example,
dynamic, adaptable and smart user interface that learns
from and with user (adaptable to different applications,
easy to use, more visual and natural with appropriate form
factor, and multi-lingual support),
ability to accept user input in many forms including voice,
ability to display rich and usable contents,
location awareness and ability to track users, products and
devices,

ability to work with and adapt to mobile commerce applications with diverse requirements (such as the types of
transaction both push and pull, transaction rates, duration,
and multicast),
possible support for context awareness,
ability to discover and download upgraded applications
and software proactively,
an operating system that can manage resources to support
many of these functions.
Some of these features are already available in hand-held
devices. For example, iMode devices using I-appli can use
software applications downloaded from sites by employing
an automatic transmission mode. These devices are also capable of receiving area specific (location-dependent) information and currently support a bi-lingual (Japanese and English)
interface (www.nttdocomo.com).
Many of these capabilities will increase the size and weight
significantly, and thus, potentially affect the usability and
portability of these devices. Additional effort is required to
better understand the engineering tradeoffs between device
capability and device usability.
The above issues are likely to affect the widespread use of
mobile commerce applications. If user interfaces are hard to
deal with, a user may not see significant value in mobile commerce applications. If out of context information is pushed
to users, many of them may not find the mobile commerce
experience to be very pleasant. If location-based services are
not usable in certain locations, it may affect whether users
adopt such applications. Clearly, privacy and security issues
are also very important, as most users do not want private information to be released to other parties. Finally, because of
the financial value of many mobile commerce applications,
security at a device level is also important. This issue becomes even more important when a user is moving between
multiple heterogeneous wireless and mobile networks. It is
possible that some of these capabilities may be realized with
the help of servers and middleware, but it should be noted
that the amount of traffic between servers and devices might
be significant.
As the number of functions available on a mobile device
increase along with improved storage and processing capabilities, an operating system that can manage the allocation
of internal resources to multiple applications and processes
is required. A general purpose OS is not suitable for such
small hand-held devices due to real-time requirements, limited processing power, memory, and screen size, and due to
the type of application that may be running. So an OS with a
small footprint and reduced storage capacity is needed. Many
emerging mobile OS are compared elsewhere [2]. These operating systems have attracted developers building applications,
which can run on handheld and other smaller devices. For

192

VARSHNEY AND VETTER

example, Microsoft has released pocket versions of its popular software such as Outlook, Excel, and Word and 3Com is
working with applications developers to create new and more
sophisticated applications for its PalmPilot. Since UNIX has
been used widely on the Internet and in other computing environments, the use of a stripped down version of UNIX with
a smaller footprint may become important for mobile applications.
Hand-held devices are likely to affect the interoperability,
implementability, and scalability of mobile commerce applications as well. Interoperability relates to the smooth interworking of m-commerce applications. A universal framework
should be followed to allow for interoperability of products
from different providers. Implementability relates to the difficulties in building m-commerce applications. As the resource
requirements of some applications may exceed the resources
provided in the devices and networks, resource constraints
need to be carefully considered. Scalability is also an issue.
As the number of users, transactions, and devices increase
the system must be able to accommodate the corresponding
increase in load and competition for computing and communication resources.
3.1. Wireless and mobile middleware for mobile commerce
Traditionally, middleware unites different applications, tools,
networks and technologies; allowing user access via a common interface. Mobile middleware can be defined as an enabling layer of software that is used by applications developers to connect their m-commerce applications with different
networks and operating systems without introducing mobility
awareness in the applications (figure 9).
The use of middleware allows applications to run with better response times and much greater reliability. Typically,
network-enabled middleware uses optimization techniques,
such as header compression, delayed acknowledgements, and
concatenation of several smaller packets into one, to reduce
the amount of traffic on the wireless networks. Some middlewares also support intelligent restarts where after a lost connection, information transfer starts at the break point and not
all over again. An example of mobile middleware is ExpressQ
from Nettech (www.nettechRF.com). It stores messages
when mobile user is out of the network range and forwards
them when mobile user comes within range the next time.

In mobile commerce environment, middleware becomes

very important due to the potential values of many mcommerce transactions, limited capabilities of devices, and
diverse set of requirements presented by various m-commerce
applications. The middleware may also affect the amount of
processing required at a device, the amount of information
that has to be transmitted over wireless networks and the level
of security for mobile commerce transactions. Also the middleware should have transaction-orientedness as needed for
many mobile commerce applications. Also, as many applications may require that certain information be sent to user
periodically or when a certain event has occurred. For these
applications, a middleware must also support push type of
information transfer to devices. As middleware may be required to support wireless Internet or adapting standard web
contents to wireless users. A possible alternative is the Wireless Application Protocol (www.wapforum.com). WAP is
designed for interoperability of different wireless networks,
devices, and applications using a common set of applications
and protocols. Using the WAP architecture, wireless middleware can be deployed as a client on a mobile terminal and as a
server on the gateway or intermediate server (figure 10). WAP
uses a micro browser as the client software and supports text,
graphics, and standard web content. It will support streaming
media such as video in the near future. In WAP, a gateway
acts as a proxy server to a mobile client and translates requests from WAP protocol stacks to protocol stacks employed
by the information server on the other side. Encoders translate the content coming from the server into compact formats
to reduce the amount of data that must be transmitted over the
wireless network. This infrastructure ensures that a mobile
user can access a wide variety of content and applications and
also allows application developers to build m-commerce applications (using proven and existing technologies) that can
run on a large base of mobile terminals.
WAP 2.0, released recently, adds the support for push
operation, allowing information to be sent without an explicit
request to users. It also supports a variety of user interfaces
and standard Internet protocols such as TCP/IP and HTTP.
It uses WML2 based on XHTML, and thus, it does not require proxy or gateway as shown in figure 10. However, to
implement the push operation, and to get improved services
and optimized communications, a WAP proxy is still neces-

Figure 9. Mobile middleware for application and content adaptation.

MOBILE COMMERCE

193

Figure 10. WAP architecture.

sary. So, the above architecture shown in figure 10 remains

valid for mobile commerce transactions.
To allow for web content to be accessible from everywhere, from PCs to TVs to Palm devices to cellular phones,
the World Wide Web Consortium (W3C) had developed several open recommendations. These recommendations include
the Extensible Markup Language (XML) for richer semantic
information, improved Cascading Style Sheets (CSS) and Extensible Style Sheet Language (XSL) to further separate content from presentation, and a Document Object Model (DOM)
which defines a language-independent application programming interface that applications can use to access and modify
the structure, content, and style of HTML and XML documents. Many of these specifications have been included in
WAP 2.0 including the support for inline and external type
style sheets, commonly supported by most Internet browsers.
Additionally, it supports long-lived sessions that may be very
useful of mobile commerce applications of longer duration.
The suspendresume function may also be used for applications that may not need to move continuous information
throughout the session. WAP 2.0 is also likely to be used in
iMode services which so far has been using a modified and
compact HTML to provide content to its wireless users.
4. Wireless networking infrastructure
In addition to mobile devices and middleware, networking
support from wireless networks is crucial in realizing mobile
commerce applications. There have been significant advances
in wireless and mobile networks in the last few years in terms
of protocols, standards, technologies, Quality-of-Service, and
user acceptance [13]. In this section, we present and discuss the wireless networking requirements for various mobile
commerce applications.
Mobile financial applications require locating a device,
user, or building in relation to the current location of a user
for making financial transactions. Due to the potential value
of such transactions, very high level of wireless infrastructure dependability may be required. To allow a user to make
these transactions from anywhere anytime, the ability to roam
across multiple wireless networks may be very useful. Similarly, many other m-commerce applications such as mobile
advertising, mobile inventory management, product locating,
and proactive service management may require location management support. Mobile advertising needs asymmetric multicast to send messages to certain users in certain locations,

but such multicast does not have to be real-time. Also for

mobile advertising, network dependability requirements may
not be stringent as loss of some messages could be tolerated.
For many m-commerce applications such as mobile entertainment and distance education services, the required Qualityof-Service in terms of bandwidth and delay may be significant. The specific requirements of each of the m-commerce
applications are shown in table 2.
From the above discussion, it appears that mobile commerce applications would present five general networking requirements: location management, multicast support, network dependability, support for Quality-of-Service, and the
ability to roam across multiple wireless networks. To help
network designers and developers, we translate these five networking requirements into more specific attributes as shown
in table 3.
Location management has attracted significant attention
in wireless community. One such reason is an FCC ruling
requiring cellular and PCS carriers to provide 911 centers
with precise location information (say, within 100 meters) of
callers (also termed E911) by 20012002 [5].
Possible solutions include the use of global positioning
satellite systems, use of several base stations for triangulation
for extracting the location information, and mixed handsetnetwork protocols for location determination [4]. Such location support is crucial for many mobile commerce applications. Fortunately, location management has seen substantial
work in the form of location-based services [6], specific location management issues [14,15], location-aware services [11],
and location- and context-aware applications [1], location architectures for B2B mobile commerce environment [3]. Other
networking requirements: support for multicast, wireless dependability, and roaming support across multiple wireless
networks have not been addressed by researchers in sufficient detail. Some early work on these issues can be found
in [12,1618].
After identifying various networking requirements and
specific attributes, we now present a global and comprehensive wireless infrastructure for mobile commerce applications
as shown in figure 11. A mobile user can use mobile commerce applications in several different ways. It could be
through infrastructure-based wireless networks such as cellular, PCS, and GSM networks, or could be through an ad hoc
wireless network that can connect to the IP-based networks
via satellites. A user could connect directly via satellite-based
systems. In this scenario, a mobile user can roam across mul-

194

VARSHNEY AND VETTER

Table 2
Specific networking requirements of m-commerce applications.
Class of applications
Mobile financial applications (B2C, B2B)

Mobile advertising (B2C)

Mobile inventory management (B2C, B2B)/

Product locating and shopping (B2C, B2B)

Proactive service management (B2C, B2B)

Mobile office (B2C)/

Wireless re-engineering (B2C, B2B)
Mobile auction or reverse auction (B2C, B2B)

Mobile entertainment services and games (B2C)/

Mobile distance education (B2C)

Wireless data center (B2C, B2B)

Wireless networking requirements

Comments

Location management
Network dependability
Roaming across multiple networks
Location management
Multicast
Roaming across multiple networks
Location management
Network dependability
Multicast
Roaming across multiple networks
Location management
Network reliability
Roaming across multiple networks
Roaming across multiple networks
Network reliability
Location management
Multicast
Location management
Network reliability
Roaming across multiple networks
QoS
Multicast
Roaming across multiple networks
Location management
QoS
Location management
Roaming across multiple networks

Secured unicast communications may

be required.
Asymmetric non-real time multicast
may be required.
Multicast preferred but series of unicast
communication may be tolerated.

Unicast or asymmetric non-real time

multicast.
Secured unicast communications may
be required
Real-time multicast with significant
user input and coordination required.

Asymmetric real-time multicast required with some user input.

Bandwidth requirements may be significant.

Note: B2C: business-to-consumer, B2B: business-to-business.

Table 3
Wireless infrastructure requirements for mobile commerce.
Networking requirements
Location management

Multicast support

Network dependability

Quality-of-Service

Roaming across multiple

networks

Specific attributes
1.
2.
3.
4.
1.
2.

Location tracking for determining the location of an object

Location accuracy and response time
Frequency of location tracking
Horizontal and vertical location tracking
Support for multicast in infrastructure wireless networks
Support for multicast in ad hoc wireless networks (much more
difficult due to dynamic topology and other factors)
3. Group connectivity under mobility/failure
4. Synchronization/atomicity of transactions from multiple users
1. Impact and frequency of component failure
2. Fault-tolerant design
3. User access to multiple networks
4. Levels of network availability
1. Bandwidth requirements
2. Delay and delay variation
3. Tolerable loss characteristics
1. Handoff among multiple wireless networks
2. Keeping track of users across networks

tiple different wireless networks and a dynamic QoS negotiation may be required due to vastly different resources and
users of these networks. This infrastructure also supports efficient location management, multicast, and provides faulttolerance to increase levels dependability. Since a user may
access multiple networks that may provide different levels of
horizontal and vertical location accuracy, response time and

transaction rate, it may be possible to select one wireless network for mobile commerce transaction by translating location
requirements of applications to specific attributes. Loss of
access to one or more wireless networks due to various failures could be overcome by switching to the remaining wireless network and still support most of the requirements of mcommerce applications.

MOBILE COMMERCE

195

Figure 11. A global comprehensive wireless infrastructure for mobile commerce.

5. M-commerce issues for carriers and developers

Wireless carriers can play a very active and important role
in the mobile commerce applications and services due to the
fact that a mobile user is going through their networks to perform all mobile commerce transactions. Additionally, a mobile user is likely to prefer one common bill (bundled services) for voice, data, and mobile commerce services. Service providers can also act as content aggregators, but are
unlikely to act as either an application or content provider
due to their focus on the networking and service aspects of
m-commerce. A service provider can also act as a clearing
house for content and application providers in advertising and
distributing their products to its customers. However, there
are many technical and non-technical hurdles (such as pricing
for mobile commerce transactions) that need to be overcome
before carriers can become major players in this emerging
field.
Currently there are many different standards for accessing wireless networks. The standards used in cellular and
PCS systems in US and Europe differ considerably. This
directly affects the interoperability and global roaming of
mobile users. This could be a major hurdle for some mobile commerce transactions involving multiple wireless net-

works. Differences in standards along with other interests

have also led to delays in the deployment of next (third) generation of wireless/mobile systems [7]. However, once 3G
networks are deployed worldwide, global roaming should become a reality. For the time being, mobile users can use
multi-function and multi-band phones to roam across several
continents (however, roaming charges do apply!). We should
note here that worldwide migration to 3G is still dependent
on operators perception of market needs, possible incentives
to carriers and operators, recovery on investment made in the
existing first and second generation wireless systems, and perceived threats to monopoly wireless carriers in many countries.
Besides the above technological issues, wireless carriers
are also likely to face challenges involving how to price
mobile commerce services, and because several carriers are
likely to be involved in completing a mobile commerce transaction, another issue is how to divide revenues among multiple carriers. There are many important issues that need to
be addressed before mobile commerce applications can be
widely deployed. These include the development of new business models for charging wireless customers and for revenue
division among providers, maturity of application software,
middleware support, vendor support, and user trust necessary

196

VARSHNEY AND VETTER

Table 4
Possible solutions to pricing of mobile commerce services and division of revenues.
Issue

Possible solutions

Comments

Pricing of mobile commerce

services

Flat rate pricing

Simple to implement
Unfair to some customers
Based on the connection time or session duration
Simple to implement
Unfair to some customers
Delays in fixed networks and web servers will affect the users cost
Based on the number of transactions attempted
Difficult to implement
Web servers and wireless links can have major affect
Based on the number of transactions completed
Difficult to implement
Desirable for most users
Based on the number of users involved in a transac- Simple to implement
tion (group multicast pricing)
Based on the number of networks/ISP/carriers in- Simple to implement
volved in a transactions
A version is currently used for roaming charges
Based on the number of packets transmitted/amount Need to count packets (even when they multiply over certain links)
of traffic
Based on the QoS received
Difficult as link quality and other factors have to be measured and
monitored throughout the session
Very fair pricing
Based on the combination of session duration, partic- More accurate pricing
ipants, number of transactions, QoS, and traffic
More complex to implement
Time sensitive pricing
Simple to implement

Division of revenues among

multiple service providers

Ratio of distances traveled by packets in individual

network
Using number of packets and number of copies made
in an individual network
Using number of switches used (or processing power)
Using number of links involved in individual networks
Using number of customers of a network involved in
the session
Using a combination of above factors

for conducting mobile transactions. Table 4 presents several

solutions to pricing and revenue division in mobile commerce
applications and services.
Now we look at issues important for developers of mcommerce applications. These issues are presented in table 5. One major issue is how to achieve application independence from the various device and wireless access technologies. However, designing for application independence
is likely to increase the actual program code as these applications can no longer take advantage of certain inherent capabilities in specific devices or networks. While developing such programs, only certain basic functionalities
about these technologies can be assumed. All of this increases the application usability across networks and devices,
but the implementation is not likely to be the most efficient.
Also as users may go through phases of intermittent connectivity or frequent disconnectivity, any impact of these
events on mobile commerce transactions should be carefully
investigated. Due to the potential values of many mobile commerce applications, atomic transactions may be necessary. It
is possible that mobile middleware may provide most of such
functions and thus reducing the amount of work needed to
support atomic transactions by applications.

Difficult to implement
Needs to count packets (even when they multiply over certain links)
Requires counting the number of switches for every packet
Requires counting the number of links for every packet
Easy to implement
More accurate but difficult to implement for revenue sharing

Table 5
Some important issues for application developers.
Issues

Comments

Network
processing
and storage
requirements

Bandwidth and delay requirements (real-time vs nonreal time applications)

Mobile device capabilities
Disconnected operation
Multicasting for group communications
Symmetric vs asymmetric processing and storage
Use of any existing Software Development Kit (SDK)
Simulation of environment where application will be
used
Maximum number of simultaneous users
The size of application code
Support for secure transactions
Support for fixed users
Independence from the underlying wireless access
technologies
Independence from the device functionalities
Interoperability with IP
Compatibility with Wireless Application Protocol
(WAP)
Support for intermittent connectivity
Adaptive to the user and network environment
Support for atomic transactions
Easy upgradability
User specified features invocation

Application
development

Compatibility
and interoperability

Desirable
features

MOBILE COMMERCE

197

Table 6
Several interesting research problems.
Research problem
1

6
7

9
10

What are some inherent differences in e-commerce and m-commerce? Do these differences require some minor modifications in
the existing e-commerce approaches or do we need to start with
completely new approaches? How mobility affects procurement,
supply chain, B2C, B2B, web site effectiveness, and value added
services?
What are some new applications that could become possible due to
the use of mobile devices and wireless networks? How the support
for these applications affects existing e-commerce implementation?
How the differences among wireless standards would affect the offering and acceptance of m-commerce applications? What are the
roles of cultural and social factors in m-commerce adoption?
What is the current level of understanding of wireless networks, devices, middleware, and applications among IS managers, systems
architects, developers, and academic researchers? What are some
ways to enhance such understanding in a shortest possible time?
What are the major factors that can affect the adoption of mobile
commerce research? Can we derive these from adoption factors of
e-commerce or from wireless networks?
What is the current level of trust in mobile commerce transactions
and how can it be increased?
What additional security problems are introduced by mobility of
users and wireless links for mobile commerce applications? Are existing network security mechanisms sufficient for mobile commerce
applications?
How pricing models for mobile commerce would be different from
e-commerce? Can same models be used in Europe, America and
other continents/countries?
What are the ways to design user interface for mobile commerce
applications that could satisfy several conflicting requirements?
What strategies carriers, vendors, providers, and managers in the
development of m-commerce applications and services should use?

6. Interesting research issues in mobile commerce

There are many interesting research problems in mobile commerce. Some of these are unique due to the limitations of
mobile devices and wireless networks and some are similar
to research problems that are currently being addressed by ecommerce researchers and developers. The problems that are
unique to m-commerce are
novel applications and services made possible due to the
wireless networks and mobile devices,
security and privacy problems that are unique to wireless
networks and mobile devices,
middleware issues that are unique due to device, network,
and protocol limitations,
role of different wireless networking standards,
adoption factors of mobile devices that are significantly
different in different parts of the world,
context and location-awareness is unique to mobile commerce as many of the applications are sensitive to the context and the location of a user.
The research problems that can be addressed by the existing e-commerce research with some modifications and extensions are

strategy of new service offering,

role of m-commerce providers,
trust building,
adoption of new services,
pricing models and sensitivity analysis.
These problems are presented in table 6. We are pursuing some of these research problems and work will be presented in future articles. It is our hope that current and new
researchers would find these problems to be of considerable
research interest and our work will fuel advances in mobile
commerce applications, adoption, networking, pricing, and
other emerging areas.
7. Conclusions and further research
Mobile commerce is an interesting and challenging area of
research and development. It presents many issues that cover
many disciplines and may best be addressed by an active participation of computer and telecommunications experts, social scientists, economists, and business strategists. In this
paper, we have looked into several issues that we consider
important for mobile commerce realization. We introduced
several new classes of applications, reviewed networking requirements, and discussed application development support.
Since the area of mobile commerce is very new and still
emerging, we have added several interesting research problems that are either currently being addressed or should be
addressed by the research and development community. We
also believe that user trust will play a crucial role in acceptance and widespread deployment of mobile commerce applications.
References
[1] C. Bisdikian, J. Christensen, J. Davis II, M.R. Ebling, G. Hunt,
W. Jerome, H. Lei, S. Maes and D. Sow, Enabling location based applications, in: Proc. First ACM International Workshop on Mobile Commerce (July 2001).
[2] R. Comerford, Pocket computers ignite OS battle, IEEE Spectrum
(May 1998).
[3] K. Cousins and U. Varshney, A product location framework for mobile
commerce environment, in: Proc. First ACM International Workshop
on Mobile Commerce (July 2001).
[4] G. Djuknic and R. Richton, Geolocation and assisted GPS, IEEE Computer (February 2001).
[5] FCC e911, http://www.fcc.gov/e911
[6] E. Gabber and A. Wool, On location restricted services, IEEE Network
(November/December 1999).
[7] IMT-2000, http://www.itu.org/imt
[8] R. Kalakota and M. Robinson, e-Business: Roadmap for Success (Addison Wesley, 1999).
[9] F. Muller-Veerse, Mobile commerce report, Durlacher Corporation, London, http://www.durlacher.com/downloads/
mcomreport.pdf
[10] M. Oliphant, The mobile phone meets the Internet, IEEE Spectrum
(August 1999).
[11] O. Ratsimor, V. Korolev, A. Joshi and T. Finin, Agents2Go: An infrastructure for location-dependent service discovery in the mobile

198

[12]

[13]
[14]

[15]

[16]
[17]

[18]

VARSHNEY AND VETTER

electronic commerce environment, in: Proc. First ACM International

Workshop on Mobile Commerce (July 2001).
U. Varshney, Supporting multicast communications over wireless and
mobile networks, A tutorial, in: ACM International Conference on Mobile Computing and Networking (MobiCom) (August 1999) (available
from the author).
U. Varshney, Recent advances in wireless networking, IEEE Computer
(June 2000).
U. Varshney, Location management support for mobile commerce applications, in: Proc. First ACM International Workshop on Mobile
Commerce (July 2001).
U. Varshney, Mobile commerce: Applications and technologies, A tutorial, in: ACM International Conference on Mobile Computing and
Networking (MobiCom) (July 2001) (slides available from the author).
U. Varshney, Multicast communications over wireless networks, Communications of the ACM (to appear).
U. Varshney, A. Snow and A. Malloy, Designing reliable, available,
and survivable wireless networks, in: Proc. IEEE International Wireless Communications and Networking Conference (WCNC) (September
1999).
U. Varshney and R. Vetter, Emerging wireless and mobile networks,
Communications of the (ACM) (June 2000).

Upkar Varshney is an Assistant Professor in the Department of Computer Information Systems at Georgia State University, Atlanta. He received a Bachelor
of Engineering in electrical engineering with Honors
from University of Roorkee, MS in computer science and Ph.D. in telecommunications & networking, from the University of Missouri-Kansas City in
1988, 1992, and 1995, respectively. Before joining
GSU, he worked as a Research Associate at the Center for Telecomputing Research (funded by Sprint,
NorTel, BNR, MCI and State of Missouri) during 19921994 and taught at
Washburn University, Kansas, during 19941998. Prof. Varshney has written
over 50 papers in mobile commerce, mobile and wireless networking, wireless dependability, and high-speed networking. He has presented some very
well received tutorials and workshops at major international conferences in-

cluding ACM MobiCom, IEEE WCNC, IFIP HPN, and HICSS. He is also
a respected speaker in wireless and mobile networking and applications. He
has delivered over 50 invited speeches, including several keynotes at major
conferences and workshops. He is actively involved with International Executive education programs at GSU. In October 2000, Prof. Varshney was
awarded the prestigious Myron T. Greene Outstanding Teaching Award. He
is on the editorial board of IEEE Computer. He has also served on the program committees of IEEE WCNC, IEEE LCN, ACM Workshop on Mobile
Commerce, HICSS and several other international conferences. He is a member of ACM and IEEE and several SIGs and societies.
E-mail: [email protected]

Ron Vetter is a Professor and Chair of the Department of Computer Science at the University of North
Carolina at Wilmington. His research interests include wireless and mobile networks, multimedia systems, net-centric computing, and web-based instruction. His research activities have been supported by
the USWEST Foundation, NSF, USDA, NASA, and
others. He received the B.S. and M.S. degrees in
computer science from North Dakota State University, Fargo, in 1984 and 1986, respectively, and the
Ph.D. degree in computer science from the University of Minnesota, Minneapolis, in 1992. Dr. Vetter serves on the program committees for several computer science-related workshops and conferences each year and publishes scholarly work on a regular basis. Since 1992, he has published more
than 70 journal, conference, and technical papers. In addition, he has received external grant funding in excess of two million dollars. Dr. Vetter
serves on the editorial advisory board for Communications of the ACM and
is an ACM Distinguished Lecturer. He also serves on the editorial board of
IEEE Computer Magazine, is an associate technical editor for IEEE Communications Magazine, is a member of IEEE Computer Societys Conference and Tutorials Board, and serves as a chair of the Net-Centric Computing focus group for the Joint Task Force on Year 2001 Model Curricula for
Computing: CC-2001 organized by IEEE Computer Society and the ACM.
Dr. Vetter is a frequent consultant to business and industry on computer communication and Internet related topics.
E-mail: [email protected]