Location Management for Mobile Commerce

Applications in Wireless Internet Environment

UPKAR VARSHNEY
Georgia State University

With recent advances in devices, middleware, applications and networking infrastructure, the wire-
less Internet is becoming a reality. We believe that some of the major drivers of the wireless In-
ternet will be emerging mobile applications such as mobile commerce. Although many of these
are futuristic, some applications including user-and location-specific mobile advertising, location-
based services, and mobile financial services are beginning to be commercialized. Mobile commerce
applications present several interesting and complex challenges including location management of
products, services, devices, and people. Further, these applications have fairly diverse requirements
from the underlying wireless infrastructure in terms of location accuracy, response time, multicast
support, transaction frequency and duration, and dependability. Therefore, research is necessary to
address these important and complex challenges. In this article, we present an integrated location
management architecture to support the diverse location requirements of m-commerce applica-
tions. The proposed architecture is capable of supporting a range of location accuracies, wider net-
work coverage, wireless multicast, and infrastructure dependability for m-commerce applications.
The proposed architecture can also support several emerging mobile applications. Additionally,
several interesting research problems and directions in location management for wireless Internet
applications are presented and discussed.
Categories and Subject Descriptors: C.2.1 [Computer-Communication Networks]: Network
Architecture and Design—wireless communication
General Terms: Applications, Performance
Additional Key Words and Phrases: Mobile commerce, location management, wireless Internet,
wireless LANs, satellites, wireless multicast, mobile applications, infrastructure dependability

1. INTRODUCTION
The wireless Internet has received significant interest among carriers, vendors,
applications developers and users. In addition to wireless access to the Internet,
it must deal with the adaptation of protocols and applications to the limitations
of devices and wireless networks [Goodman 2000]. Although several different

This research was supported, in part, by a research grant from the Robinson College of Business
of Georgia State University.
Author’s addresses: Department of Computer Information Systems, Georgia State University,
Atlanta, GA 30302-4015; email: [email protected].
Permission to make digital or hard copies of part or all of this work for personal or classroom use is
granted without fee provided that copies are not made or distributed for profit or direct commercial
advantage and that copies show this notice on the first page or initial screen of a display along
with the full citation. Copyrights for components of this work owned by others than ACM must be
honored. Abstracting with credit is permitted. To copy otherwise, to republish, to post on servers,
to redistribute to lists, or to use any component of this work in other works requires prior specific
permission and/or a fee. Permissions may be requested from Publications Dept., ACM, Inc., 1515
Broadway, New York, NY 10036 USA, fax: +1 (212) 869-0481, or [email protected].
°
C 2003 ACM 1533-5399/03/0800-0236 $5.00

ACM Transactions on Internet Technology, Vol. 3, No. 3, August 2003, Pages 236–255.
 Location Management for Wireless Mobile Commerce Applications • 237

Fig. 1. A possible architecture for the wireless Internet.

configurations are possible, we believe that the wireless Internet should in-
clude devices and user interfaces, wireless infrastructure, mobile middleware,
servers, and databases (Figure 1). In such an environment, a typical transac-
tion consists of multiple steps involving several components. A user’s request is
forwarded to a transaction server, which checks with an authentication server.
Once authenticated, preference and pricing databases are accessed. Also, to
locate other users and servers, a location database is contacted. If the transac-
tion requires setting up a group session involving multiple users, then a mul-
ticast server is also contacted. Certain transactions require accessing a data
center and additional third party services, such as location tracking by other
providers.
Although the architecture shown in Figure 1 corresponds to an
infrastructure-oriented wireless network, the transactions could also be sup-
ported in ad hoc wireless networks, where users, devices, and servers could all
be mobile. One or more nodes (devices or servers) can dynamically be selected
to act as group leaders based on mobility levels and their ability to locate other
required servers. In an ad hoc wireless network, completing a complex and pos-
sibly long session before a component moves out of range is a challenging task
and certainly more work then is necessary to address transactions in ad hoc
wireless networks.
We believe that one of the major drivers of the wireless Internet will be
mobile commerce, which includes many new applications that become possible
only due to wireless infrastructure (such as location-based services) in addition
to e-commerce applications modified for wireless environment. It is estimated
that the market of several hundred billion dollars for m-commerce will exist in
the next few years.
As mobile commerce is an emerging area, there has been limited work in
addressing a range of complex issues. This includes a classification of applica-
tions [Varshney et al. 2000], a 4-layer framework and a detailed treatment of
m-commerce issues [Varshney and Vetter 2002], and location issues and imple-
mentation experiences [Cousins and Varshney 2001, Varshney 2001].
ACM Transactions on Internet Technology, Vol. 3, No. 3, August 2003.
238 • Upkar Varshney

Although there are several possible m-commerce applications [Varshney and

Vetter 2002], only a few of these applications have been offered by providers
thus far. These include preliminary versions of mobile financial services, mo-
bile advertising, and location-aware services. Advanced versions of these ap-
plications will require location information of users, devices, servers, products,
and services. These applications also have widely different location precision,
response time, and scalability requirements. Also, more work is necessary to
evaluate location overhead under different mobility patterns, transaction rates,
and network size. There has been very little work in this area, but we would
like to briefly overview the related work in location-based services.
An interesting implementation of location-based discovery of services in-
cludes mapping current location of a user to an area in Cellular Packet Digital
Data (CDPD) and then searching services in the area [Rastimor at el. 2001]. An-
other service discovery, based on mapping of bookmarks to Universal Resource
Locator (URL) of services in the current location of user, is presented in Sastry
et al. [2001]. A secure location service using context and location awareness is
presented in Bisdikian et al. [2001].
In this article, we present an integrated location management architecture
that has been designed to support the diverse location requirements of both
current as well as future m-commerce applications. This proposed architec-
ture is capable of supporting a range of location accuracies, network cover-
age, and, wireless multicast and dependability requirements of m-commerce
applications.
Major contributions of this article are (a) requirement analysis of location-
intensive mobile commerce applications and (b) design of an integrated location
management architecture for m-commerce applications. We have also identi-
fied several interesting research problems, directions, and possible solutions in
location management for m-commerce.
The article is organized as follows: In Section 2, we discuss three m-commerce
applications and specific location requirements. In Section 3, we present our
location management architecture. In Section 4, we show how the proposed
architecture may be applied to support future m-commerce applications. We
present several research issues and future directions in Section 5 and some
concluding remarks in Section 6.

2. LOCATION-INTENSIVE M-COMMERCE
APPLICATIONS AND REQUIREMENTS
Several m-commerce applications have been proposed in the literature
[Varshney and Vetter 2002]; however, only some of these have been offered by
wireless providers so far. In this section, we discuss three current m-commerce
applications that require location support. These location-intensive applica-
tions are:

—Mobile financial applications

—Mobile advertising
—Location-based services.
ACM Transactions on Internet Technology, Vol. 3, No. 3, August 2003.
 Location Management for Wireless Mobile Commerce Applications • 239

Fig. 2. Mobile financial applications.

2.1 Mobile Financial Applications

Mobile Financial Applications consist of mobile banking and brokerage services,
mobile money transfer, and mobile payments. These applications could trans-
form a mobile device into a business tool, replacing bank, ATM, and credit cards
by letting a user conduct financial transactions with mobile money. It is pro-
jected that the number of users making mobile payments will reach to a total of
285 million in Western Europe, Asia and North America by 2005. Out of these,
about 53 million users are expected to have a mobile wallet [Allnetdevices,
2002].
Several mobile financial applications are shown in Figure 2. A mobile user
attempts to purchase goods or services from a business or service provider,
which then contacts a trusted third party, the wireless service provider, or a
financial institution to authenticate the user and amount of purchase. Once
approved, a mobile payment can be made and the purchase is completed. The
corresponding funds can then be withdrawn from user’s m-wallet, charged to
user’s phone bill, or subtracted from user’s bank account. Alternatively, the user
could pay using “mobile money” provided to him/her by another user or a third
party mobile money provider. Mobile money can be moved freely among users
either by using a local area wireless network or by using the wireless service
provider’s network.
Two other requirements of mobile financial applications are support for mo-
bile payments and secure transactions. Several groups are already working
on mobile payments, including PayCircle that is established by HP, Lucent,
Oracle, Sun, and Siemens [Varshney 2002]. Mobile payments can be divided
into macro-(more than $10) and micro-payments (less than or equal to $10).
Micro-payments can be implemented in several ways: first, a user could make
a toll call equal to the cost of item, as used by SONERA in Finland for wire-
less vending service. Second is to buy prepaid numbers from a service provider,
bank, or credit card company, as used in PayPal for mobile money transfer. For
mobile payment providers, there are issues, such as the transaction cost of mo-
bile micro-payments, and ways to make some profit on mobile micro-payments.
ACM Transactions on Internet Technology, Vol. 3, No. 3, August 2003.
240 • Upkar Varshney

First, service providers could require prepayments by users. Also, the cost
of micro-payments decreases with an increased number of transactions. A
provider can also offer mobile payment service as a competitive feature, re-
sulting in an increase in customers and overall revenues.
There are several mobile payment providers, including wireless provider
Vodafone’s service in England, Germany, and Italy [Vodafone 2002]. Expected
to reach to all of its 50 million customers by late 2003, the mobile payment
service is based on user and device authentication, and e-wallet processing.
Another payment service to businesses and customers has been started by Pay-
box using its own proprietary architecture [Paybox 2002]. To support mobile
payments across multiple operators, Sprint and eONE have initiated a mobile
payment network that will allow users to make mobile payments nationwide
[Sprint 2002]. By using PaymentWorks, the same software used by several
providers in Europe, the network can even allow international payments. Sev-
eral financial companies and vendors are implementing end-to-end transaction
support for financial applications [FSTC 2002]. Some current work on mobile
cash withdrawal and change return can be found in Tracz and Wrona [2001].
Mobile financial transactions require a strong level of security support. Al-
though, security features have been added in mobile middleware such as WAP,
end-to-end security is still a problem [Ghosh and Swaminatha 2001]. End-
to-end security support will become possible with the wide scale deployment of
WAP 2.0. For financial applications, wireless PKI (Public Key Infrastructure), a
system to manage keys and certificates, can be used to authenticate and obtain
digital signatures from mobile users. In Japan, a few simple mobile financial
applications are currently supported for iMode phones using KVM, a version
of Java designed for small devices. Security is provided by SSL (Secure Socket
Layer), both 40 and 128 bit versions [Varshney 2002].

2.2 Mobile Advertising

By keeping track of user’s purchasing habits and current location, a very tar-
geted advertising campaign can be performed. In one possible scenario (shown
in Figure 3), mobile users are informed about various on-going specials. Mes-
sages can be sent to all users who are currently in a certain area (identified
by advertisers or even by users) or to certain users in all locations. Depending
on interests and personality types of individual users, advertisers could decide
whether a “push” or “pull” form of advertising is more suitable. As more wire-
less bandwidth becomes available, content-rich advertising can be produced for
individual users with specific needs, interests, and inclinations [Varshney and
Vetter 2002]. It has been demonstrated in several trials that mobile users are
willing to receive advertising messages with incentives.
From wireless service provider’s view, several factors including the size of
advertising messages, number of customers targeted and the value of items can
be used to determine how to charge for mobile advertising. It should be noted
that there are major issues of privacy and sharing of user information with
other providers. We believe that an “opt-in” approach should be implemented
where explicit user permission is obtained before “pushing” any advertising
ACM Transactions on Internet Technology, Vol. 3, No. 3, August 2003.
 Location Management for Wireless Mobile Commerce Applications • 241

Fig. 3. A possible scenario for mobile advertising.

content. Cellular Telephony and Internet Association (CTIA) has requested

that the FCC approve its “opt-in” policy for mobile advertising.
There has been some progress on mobile advertising. Mobile advertisements
with coupons for a fast food chain have been tested by SMS-Rabatt [Fichtner
2002]. It has been considered quite successful as more than 3.5% of coupons
were utilized as lunch-time messages were sent to interested users, resulting
in very targeted and timed mobile advertising. Pepsi Co. has created an adver-
tisement using “Pepsi Foot”, a game where mobile users select a soccer team to
play with other virtual teams [Diercks 2001]. Many wireless carriers including
AT&T wireless, BT Cellnet, and Vodafone are offering mobile advertisements
using opt-in approach. In Japan, several portals are pushing simple advertis-
ing contents to iMode phone users. To create a uniform set of guidelines for
wide scale use of mobile advertising, the Wireless Advertising Association and
Wireless Marketing Association have merged, thereby bringing many players in
mobile advertising together including Unilever, AOL, Nokia, Orange, Vodafone,
and Lycos [Mobile Marketing Association 2001].

2.3 Location-based Services

These services utilize a user’s location to provide location-aware content includ-
ing information on restaurants, devices, users, and products. Location-based
services can be offered in both “pull” and “push” modes. For example, one user
might be interested in knowing the availability and waiting time at one or more
restaurants close to her current location (pull). Another user might like to be
informed when one of his friends is located in the same general area (push).
In general, these services require location tracking of fixed, portable and mo-
bile entities. Location information of all fixed entities can be kept in a separate
database for each area, while location tracking of mobile and portable entities
could be performed on-demand. When a mobile user enters an area, the list of
services and location information can be provided based on current preferences
and/or the history of choices.
ACM Transactions on Internet Technology, Vol. 3, No. 3, August 2003.
242 • Upkar Varshney

Fig. 4. Personalized location-based services.

A personalized location-based service architecture is shown in Figure 4. In

this scenario, once a user enters a designated area, the user information from
previous networks and locations will be accessed. The system will also deter-
mine what types of location-aware services this user has subscribed to or is
authorized to access.
From an implementation point of view, several factors to consider are: (a) how
to charge businesses for providing their location information to mobile users,
(b) how to ensure correctness and usability of location information, (c) how to
charge customers, and, (d) how to divide revenues among multiple wireless,
location, and database providers.
Currently, there are a few examples of location-based services, not neces-
sarily personalized or user-specific. Several location-based services are being
offered by carriers such as mapping, routing, and list of places. Although not all
of these offerings have become an instant success, there is considerable inter-
est among users and businesses as several pricing options are offered. Vendors
are offering to wireless service providers several fully functional location-based
systems, including a Nortel system that is being deployed in several countries
to support location-aware services to users and businesses [Nortel 2002]. There
has also been some research and development work in location-based services.
The location of a user is mapped to an area and then using Cellular Packet
Digital Data (CDPD) base station numbering, services are searched in the area
[Rastimor et al. 2001]. A service discovery, where bookmarks are mapped to
Universal Resource Locator (URL) of services in the current location, has been
implemented [Sastry at el. 2001]. Location changes are detected by applica-
tions and URLs are dynamically mapped. A secure location service has been
implemented by using context and location awareness [Bisdikian et al. 2001].

2.4 Specific Location Management Requirements of M-commerce Applications

Location management requirements of m-commerce applications can be ex-
pressed in terms of location accuracy, response time and priority, network cov-
erage, number of devices involved, wireless dependability and reliability, type
of communications, number of location queries per transaction, and, frequency
ACM Transactions on Internet Technology, Vol. 3, No. 3, August 2003.
 Location Management for Wireless Mobile Commerce Applications • 243

and duration of transactions. Location accuracy depends on the radio char-

acteristics, response time needed, level of mobility, network topology, and the
positioning scheme used. Some applications require both horizontal as well as
vertical positioning, speed and direction, and a certain confidence level with
a location estimate. The location requirement is also likely to be time- and
location-dependent.
Failures in a wireless infrastructure will greatly affect m-commerce applica-
tions. Failure of Home Location Register/Visitor Location Register (HLR/VLR)
that stores approximate locations of wireless users will affect location-intensive
applications [Varshney and Malloy 2001]. Location management should cover
both indoor and outdoor tracking of people, products, and services. A user is
more likely to need location information when roaming under another provider
whose network may not have the same level of location performance. Some ap-
plications involve users and devices over several wireless networks of different
location accuracies. A “minimum capability set” of each wireless network should
be matched with minimum requirements of each m-commerce application.
Other issues in location management in mobile commerce applications are
privacy and security, ownership of location information, device complexity and
user interface, overhead of location-based services, development of business
models for pricing and revenue sharing, vendor support, and service interop-
erability. Access to current location information of users raises both security
and privacy issues. This potentially makes certain users vulnerable to security
threats. A user could specify who could access such information in “normal”
and “emergency” cases and also what and when some information can to be
“pushed” to him/her. One issue is who owns the location information. Since a
wireless carrier provides the first access point for most wireless users, it is likely
to have such information. A user should be able to specify that such information
can only be used by the wireless service provider and must not be distributed
to others. However, a variety of information on products, services, and places
could be aggregated and then location-sensitive and user-specific content can
be created.
We now discuss specific requirements of the three location-intensive applica-
tions (Table I). Mobile financial services require the location tracking of finan-
cial institutions, mobile users, transaction servers, and devices that are capable
of accepting micro-payments. Such an application would require location accu-
racy of the order of few meters. Since a few devices are likely to be involved in
any financial transactions, unicast mode of information transfer could be used.
Due to the potential value of financial transactions, dependability of wireless
infrastructure would be a critical requirement. On average, a user is likely to
make a few such transactions in a day and we believe that most users would
finish such transactions in a few seconds.
Mobile and user-specific advertising applications present very different re-
quirements, since mobile advertising needs to reach specific users served by
one or more wireless networks and the same message has to be transmitted to
multiple users. So, location management, multicast and roaming across mul-
tiple networks are required. Since these services need location information, a
third party could be used for location services. Since the potential value of such
ACM Transactions on Internet Technology, Vol. 3, No. 3, August 2003.
244 • Upkar Varshney

Table I. M-commerce Applications and Location Requirements

Personalized
Mobile Financial Mobile and Locational location-based
Applications Advertising services
Location Precision Meters (sub-cell) Hundreds of meters Meters (sub-cell)
(upper limit) (cell)
Response time and Seconds/few times Minutes/once in a Seconds/several
frequency per in a transaction transaction times in a
transaction transaction
Required Wireless Citywide Small area to Citywide to
Network Coverage citywide nationwide
Number of devices Few Several Several
and entities
involved
Information transfer Secured unicast Asymmetric Asymmetric
mode non-real-time real-time unicast
multicast or multicast
Wireless Very high Can tolerate lower High
dependability dependability
requirement
Transaction frequency Once a day for few Few times a day for Several times a day
and duration seconds few seconds for few minutes

advertisements is not as high as financial transactions, wireless dependability

is not so crucial. However, to cover a significant number of users in a given
area, as is needed in mobile advertising, wireless multicast support is very im-
portant. If such support is not available, either the number of users has to be
limited or wireless networks will have to handle significantly increased traf-
fic. On the positive side, since these advertisements could wait for seconds to
minutes, real-time delivery is not necessary. Also users do not have to be pre-
cisely located as cell-level accuracy is sufficient for mobile advertising which is
targeting users in a given area of approximately 1 mile or less.
Personalized location-based services introduce requirements that are likely
to be somewhere in between financial applications and mobile advertising. The
number of entities involved is likely to be more than that for financial appli-
cations but certainly lower than that for mobile advertising. The accuracy of
information must be high but the wireless dependability requirement is likely
to be lower than that required for financial applications. These requirements
are presented and compared in Table I.

3. AN INTEGRATED LOCATION MANAGEMENT ARCHITECTURE

Having presented m-commerce applications and their specific requirements, we
now present an integrated location management architecture. The proposed
architecture (Figure 5) consists of several heterogeneous wireless networks
including satellites, cellular, PCS and 3G networks, and, wireless LANs and
PANs to support diverse location requirements. These are location precision,
wireless coverage, multicast, and wireless dependability. The location precision
requirement can be satisfied by using one of several wireless networks, which
ACM Transactions on Internet Technology, Vol. 3, No. 3, August 2003.
 Location Management for Wireless Mobile Commerce Applications • 245

Fig. 5. The proposed integrated location management architecture.

provide different levels of location accuracy. An extensive wireless coverage is

achieved by providing indoor and outdoor coverage to fixed and mobile users
in local as well as wide area environments. Access to multiple wireless net-
works also enhances the infrastructure’s dependability. The wireless multicast
requirement is fairly complex and most current wireless networks are not de-
signed to support multicast. In our proposed architecture, wireless multicast
could be supported by satellite systems as well as by cellular 2G/3G systems
using a selective broadcast or a tree-based multicast approach [Varshney 2002].
The wireless infrastructure dependability requirement is supported by fault-
tolerant operation at device, access and infrastructure levels. This requires that
a user device be equipped with multiple network interfaces; however it provides
fault-tolerant (highly dependable) operation for both user devices and networks.
It is possible that a user is able to access only one network in a certain location
at a certain time. Thus, the level of accuracy achieved is both time and location
dependent. Other m-commerce requirements are transaction frequency and du-
ration, and response time. These are dependent on network traffic and would
be addressed in our future work dealing with location management overhead
and performance evaluation.
Since this architecture supports the roaming of a user across multiple
networks, location co-ordination is necessary among networks. As shown in
Figure 5, location tracking can also be performed using a wireless local area
network (WLAN) or a personal area network (PAN). Since these networks
cover smaller areas (and fewer users), a base station or a certain device can
ACM Transactions on Internet Technology, Vol. 3, No. 3, August 2003.
246 • Upkar Varshney

Fig. 6. Satellite-based systems for m-commerce applications.

be programmed to determine if a certain device or user is located in its cover-

age and this is applicable to both infrastructure and ad hoc versions.
Location tracking involves mobile, portable, or fixed entities. Mobile entities
can be located within the accuracy of the location scheme of the wireless net-
work(s) used. If mobile entities are part of an ad hoc wireless network, then a
more specialized scheme (such as using GPS and a monitoring system) has to
be used. Portable entities without regular wireless access can be tracked using
radio frequency tags or a specialized locator device. Location information on
fixed entities can be stored in a database and can be updated as necessary. As
shown in Figure 5, a user with access to such infrastructure can locate devices,
products, and location-aware mobile information services.
Now we describe the individual components of our proposed architecture.
These include satellite-based systems, 2G and 3G networks, WLANs and PANs.

3.1 Satellite-based Systems

Satellite-based systems can be used for wide area location tracking. One such
system is GPS (Global Positioning Satellite-system), where 24 satellites broad-
cast coded location information. The information is received and processed by
GPS receivers to determine their locations. The location precision achieved is
in the range of a few to several hundred meters. To reduce the complexity of
user devices, assisted GPS system could be used. To improve location accuracy
even further, differential GPS systems can also be deployed. In our proposed
architecture, the GPS information could be mapped to a specific area and then
used to receive location-based services as shown in Figure 6.

3.2 Cellular Wireless Networks

In cellular, PCS, and GSM, location tracking is done as follows: as long as the
user stays in a certain location area, it does not update the location. However,
once the user moves to a different location area, update messages are sent
(Figure 7). In general, the network knows the location of users with accuracy
equal to the size of the location area. Our proposed architecture achieves even
higher accuracy by using a smaller location area with a reduced cell size and
ACM Transactions on Internet Technology, Vol. 3, No. 3, August 2003.
 Location Management for Wireless Mobile Commerce Applications • 247

Fig. 7. Location tracking by current cellular wireless networks.

Fig. 8. Using small cells and base station triangulation to achieve precise location.

number of cells. The reduced number of cells allows accurate location informa-
tion to be found quickly by paging fewer cells. In case an “immediate” tracking is
needed with near-zero response time or when a user is inactive (or is in an area
out of the network coverage), the network can return the last known location
of the user.
Even higher precision of location could be supported by the emerging En-
hanced 911 (E911) infrastructure that will allow network-based tracking with
100-meter precision and handset-based tracking with 50-meter for mobile users
[FCC 2001]. The accuracy achieved for portable and fixed entities is even
higher. Major E911 schemes are Assisted and Differential Global Positioning
Systems (A-GPS and D-GPS), Time Difference of Arrival (TDOA), Angle of Ar-
rival (AOA), and Location Pattern Matching (LPM) [Djuknic and Richton 2001].
TDOA and AOA schemes locate a mobile device by processing the difference in
signal arrival times at three or more antenna sites, termed base station trian-
gulation. Our architecture supports even higher location accuracy by combining
small cells with base station triangulation as shown in Figure 8.
ACM Transactions on Internet Technology, Vol. 3, No. 3, August 2003.
248 • Upkar Varshney

Fig. 9. Designing smaller wireless LANs for precise location tracking.

3.3 WLANs, PANs and RFID

Some of the cellular/PCS location schemes can be used in indoor location track-
ing. Since many indoor applications require higher location precision, we pro-
pose that smaller wireless local area networks (WLANs) and personal area
networks (PANs) be used. In our proposed architecture, base stations are kept
closer and these co-operate in location tracking of radio-enabled devices, users,
products, and services. The radius of a cell can be determined as the minimum
of location accuracy and the coverage of base stations as shown in Figure 9.
The last component of our proposed architecture is RFID (Radio Frequency
Identification). It uses wireless links to uniquely identify objects or people us-
ing dedicated short-range communications [D’Hont 2001]. When a product or
person with a tag enters the read zone of a reader, the address and data stored
on the tag is read and can be sent to a server for location-tracking purposes.
Since RFID readers have limited coverage (5–10 feet), our proposed architec-
ture includes a multidimensional grid of RFID readers to cover the whole area
(such as a warehouse). This would detect both horizontal and vertical location
of components, products and people with RFID tags. The maximum distance
between two neighboring readers can be based on the range of readers and the
location accuracy required as shown in Figure 10.
Another interesting way to perform indoor location tracking is via using spe-
cialized location devices attached to products and people. One such example is
Locus that can be attached to clothes or portable devices [Koshima and Hoshen
2000]. Such location terminals can return signal strength data and IDs of base
stations to a monitoring center. Using this information, the user location can
be computed to a varying degree of accuracy.

4. USING THE PROPOSED ARCHITECTURE TO SUPPORT

FUTURE M-COMMERCE APPLICATIONS
In this section, we discuss how the proposed architecture could be used to sup-
port future mobile commerce applications such as product recommendation
systems, wireless re-engineering, and mobile auction. As these applications
are yet to be offered by wireless carriers or content providers, it is much harder
to derive their specific location requirements. But, in any case, we attempt to
ACM Transactions on Internet Technology, Vol. 3, No. 3, August 2003.
 Location Management for Wireless Mobile Commerce Applications • 249

Fig. 10. Designing RFID-based grid for location accuracy.

derive the general requirements of these applications and then discuss how
our proposed location management architecture is capable of supporting these
emerging and future m-commerce applications.

4.1 Future M-commerce Applications

Wireless Re-engineering includes applications that focus on improving the qual-
ity of business services using mobile devices and wireless infrastructure. An
example of such application is instant claim-payments by insurance companies
[Varshney and Vetter 2002].
Mobile Auction and Wireless Trading includes applications that allow users
to buy or sell certain items using multicast support of wireless infrastructure.
These applications require that group membership not be adversely affected by
brief wireless disconnectivity or intermittent connectivity as observed in many
wireless networks of today. An example could be airplanes competing to buy a
landing time slot during runway congestion (a proposed solution to air-traffic
congestion problem).
Product Recommendation includes applications allowing users/businesses
to receive recommendation of various products and services from a third
party/other customers. Since there is a potential for conflict-of-interest from
certain sets of users, some guidelines are necessary to create user trust before
such applications become reality.
Mobile Entertainment Services and Games (B2C ) are applications providing
entertainment services to users on a per event or subscription basis. These in-
clude video-on-demand, audio-on-demand, and interactive games. It appears
that mobile games will become one of the drivers of the wireless Internet, es-
pecially if group connectivity and response time issues can be addressed.
Mobile Office is an application providing a complete office environment to
mobile users anywhere anytime including traffic jams, airport, and conferences.
Although not as exciting as many other mobile applications, this is a useful
application for business travelers and workers on the move.
ACM Transactions on Internet Technology, Vol. 3, No. 3, August 2003.
250 • Upkar Varshney

Mobile Distance Education extends distance/virtual education support for

mobile users everywhere allowing a class using streaming audio and video.
This application would be very useful if user mobility could be supported over
one or more wireless networks to allow for continued connectivity.
Wireless Data Center application supports a large amount of stored data to
be made available to mobile users for making “intelligent” decisions as detailed
information on one or more products can be downloaded. It is an important
application that is quite appealing to managers and decision makers who have
to make certain decisions while on the move and under time pressure.
Wireless re-engineering, mobile office and wireless data center applications
have similar requirements and thus we will discuss these together. These ap-
plications require a cell level (hundreds of meters) location accuracy, very high
wireless dependability, wide coverage, and unicast operation for transactions
that involve few devices and need response time of a few seconds. Mobile auc-
tion application requires similar location accuracy and dependability; however,
it needs real-time wireless multicast involving a number of users spread over
a possibly large area. Also, the transactions will last for several minutes and
possibly longer. Mobile entertainment services and games, and mobile distance
education have similar requirements. These applications also require cell level
location accuracy, but dependability requirements are not very high. These
applications require real-time wireless multicast; however, the amount of in-
formation flow in two opposite directions are likely to be quite different. All of
these requirements can easily be met by the proposed location management
architecture, which supports various degrees of location precision, wireless de-
pendability, coverage, and multicast.

4.2 Supporting Future M-commerce Applications

We now discuss how our proposed architecture can support the above require-
ments. The diverse location precision of different m-commerce applications can
be obtained by carefully selecting and using one of the several wireless networks
as shown in Figure 5.
Outdoor tracking support for m-commerce applications may be provided by
either a cellular/PCS system or a satellite-based system such as assisted GPS.
Nearly all the schemes used in cellular and satellite-based networks do well in
an outdoor environment. Even wireless LANs and RFID-based systems could
support applications requiring outdoor location management. Many of the pro-
posed outdoor schemes will encounter performance problems in indoor envi-
ronment due to triangulation difficulties caused by weaker signals and line
of sight requirements of satellite-based schemes. Since many indoor tracking
applications require higher location precision, smaller wireless local area net-
works (WLANs) and personal area networks (PANs) should be used for indoor
location management. Indoor tracking for m-commerce could be performed us-
ing specialized cells (where a base station can locate in a very small area, but
a significant number of base stations are required to cover the whole area),
wireless LANs, ad hoc personal area networks (PANs), and, Radio Frequency
ID (RFID).
ACM Transactions on Internet Technology, Vol. 3, No. 3, August 2003.
 Location Management for Wireless Mobile Commerce Applications • 251

Fig. 11. Multicast support in the proposed architecture.

One of the requirements of m-commerce applications is wireless multicast

support. This is necessary for applications involving multiple users and enti-
ties. The current wireless infrastructure was not designed for such support;
however, our proposed location management architecture can support multi-
cast as shown in Figure 11. This architecture can support multicast in both
connection-oriented as well as connection-less mode. These modes will facili-
tate “group-oriented” m-commerce applications that may require long sessions
(such as mobile multiparty games) and short sessions (mobile advertising or
location-based services), respectively.
The service providers must ensure that users, products, and services are cov-
ered by one or more wireless networks. Such coverage will require that wireless
networks from multiple providers are interoperable. Also, in near future, a mo-
bile user will require roaming support across several different types of wireless
networks as opposed to several networks of the same type today. In such case, lo-
cation co-ordination will be necessary among heterogeneous wireless networks.
Depending on the network architecture, all location requests go to a loca-
tion server. In our proposed solution, the middleware negotiates or requests
specific requirements (precision, response time, etc.) from multiple networks
or schemes as shown in Figure 12. The middleware can translate application’s
requirements into specific location attributes such as location precision and
response time for the underlying wireless infrastructure. It can also translate
network performance attributes into specific requirement support. These capa-
bilities are required for negotiating application’s requirement and selecting a
location scheme. The protocol and steps used by the middleware are shown in
Figure 13. Also, as one or more wireless networks may have limited resources,
a trade-off between the number of users needing location tracking and possible
response time (or accuracy) would have to be made.
ACM Transactions on Internet Technology, Vol. 3, No. 3, August 2003.
252 • Upkar Varshney

Fig. 12. Selecting a location scheme or network using applications requirements.

Fig. 13. The protocol used by middleware in the location management architecture.

Middleware should also negotiate support for location management even

when the user is out of its home network while roaming to another network.
The network will have very different location accuracy, response time, location
scheme and format in which the location information is presented. To make
things more difficult, the location information requested by the roaming user
ACM Transactions on Internet Technology, Vol. 3, No. 3, August 2003.
 Location Management for Wireless Mobile Commerce Applications • 253

for its own location or for locations offering certain services and products is
more likely to be needed and is certainly more critical.

5. RESEARCH ISSUES AND FUTURE DIRECTIONS

IN LOCATION MANAGEMENT
After discussing the requirements of m-commerce applications and our pro-
posed location management architecture, we now turn to several interesting
research problems related to location management in a wireless Internet. We
will discuss application-specific design of location management schemes, com-
putation of location management overhead, scalability and related issues in
location-intensive mobile commerce applications.
The design of application-specific location management schemes could be
based on specific location requirements (such as required location accuracy
and response time) of one or more applications. Then signaling, processing
and storage requirements of location scheme could be computed. On the other
hand, locational attributes of an underlying wireless networks could be con-
sidered in deciding which of the several classes of mobile commerce applica-
tions could be supported. Applications that could not be easily supported by
a given location scheme could be modified or redesigned to work with lower
than expected performance such as reduced precision, increased delays or lim-
ited number of users. Applications could adjust to lesser attributes of the
underlying network, especially when a mobile user is roaming across multi-
ple heterogeneous wireless networks as supported by the emerging wireless
standards.
The location management overhead of m-commerce applications can be es-
timated by modeling or simulating applications in terms of location precision,
response time, transactions frequency, update rate and transaction duration,
number of entities involved and the coverage needed. Such modeling will re-
quire individual steps of m-commerce applications in both “push” or “pull” ver-
sions. In “push” version, the network or a server initiates a transaction based on
the current time (network pushes some information every hour), user’s current
location (whenever the user comes within a mile of a store, network sends an
advertising message), or a threshold using context awareness and other user
specific information.
Scalability of various location-intensive mobile commerce applications could
be determined by network resources, the maximum number of concurrent lo-
cation requests, the desired response time, and the number of transactions.
Interoperability issue will become important as some m-commerce applica-
tions involve multiple networks that provide different location performance.
In such cases, interoperability must be ensured by agreeing on “minimum com-
mon functionalities”.
As locational resource requirements of some applications could exceed net-
work resources and device capabilities, implementability issues should be care-
fully addressed. The requirements can be measured using simulation tools
and then the applications can be scaled-down or simplified to enhance im-
plementability. Such flexibility in various m-commerce applications will allow
ACM Transactions on Internet Technology, Vol. 3, No. 3, August 2003.
254 • Upkar Varshney

working with reduced locational accuracy, increased response time, reduced

reliability and other factors.
Another interesting problem is context awareness. Context awareness could
be used in deciding if certain actions (such as pushing advertising information)
should be completed at certain time. It is difficult to derive context awareness,
but it is possible to use recent user actions in deriving a context.

6. CONCLUSIONS
In this article, we have presented the requirements of several location-intensive
mobile commerce applications and an integrated location management archi-
tecture to support these requirements. The proposed architecture is shown to
be capable of supporting diverse requirements of m-commerce applications in
terms of location accuracy, wireless multicast, dependability, response time, and
wireless coverage. We also showed how the proposed location architecture could
be used to support future m-commerce applications. Our future research work
includes location co-ordination among multiple wireless networks, location ne-
gotiation protocols for m-commerce, evaluation of m-commerce locational over-
head, prioritization of location requests based on applications requirements,
context (emergency, anxiety, etc), and processing delays. Other open research
issues are interoperability, implementability, middleware support, and context
awareness. We hope that many of the above issues will be addressed in the
near future, enabling the wide scale deployment of wireless Internet applica-
tions and services.

ACKNOWLEDGMENTS
I would also like to express my deep appreciation for Prof. Won Kim, Editor-
in-Chief, TOIT for his extremely valuable feedback to early versions of this
article. Many thanks also go to Prof. Ron Vetter of University of North Carolina-
Wilmington for his suggestions to improve the readability of this article.

REFERENCES
ALLNETDEVICES. 2002. Mobile Payments Set to Soar. (www.allnetdevices.com/wireless/news/2002/
02/07/study mobile.html).
BISDIKIAN, C., CHRISTENSEN, J., DAVIS II, J., EBLING, M. R., HUNT, G., JEROME, W., LEI, H., MAES, S., AND
SOW, D. 2001. Enabling location-based applications, In Proceedings of the 1st ACM Interna-
tional Workshop on Mobile Commerce (Rome, Italy). ACM, New York, 38–42.
COUSINS, K. AND VARSHNEY, U. 2001. Location management in B2B mobile commerce environment,
In Proceedings of the 1st ACM International Workshop on Mobile Commerce (Rome, Italy, July).
ACM, New York, 43–48.
D’ HONT, S. 2001. The cutting edge of RFID technology and applications for manufacturing
and distribution. A Texas Instruments White Paper (available at http://www.ti.com/tiris/docs/
manuals/whtPapers/manuf dist.pdf)
DIERCKS, R. 2001. Mobile advertising: Not as bad as you think, Wireless Internet Mag. (July-
August) (available at www.wirelessinternetmag.com/news/0108/0108 research ads.htm).
FCC ENHANCED 911(www.fcc.gov/e911).
DJUKNIC, G. AND RICHTON, R. 2001. Geolocation and assisted GPS, IEEE Comput. 34, 2, 123–125.
FICHTNER, H., 2002. Marketing via mobile phone: Coupons conquer the mobile phone, Acquisa
(available http://www.evision-ventures.de/eng/news/029.html).
FINANCIAL SERVICES TECHNOLOGY CORPORATION (www.fstc.org).

ACM Transactions on Internet Technology, Vol. 3, No. 3, August 2003.

 Location Management for Wireless Mobile Commerce Applications • 255

GHOSH, K. A. AND SWAMINATHA, T. N. 2001. Software security and privacy risks in mobile
e-commerce. Commun. ACM 44, 2, 51–57.
GOODMAN, D. 2000. The wireless Internet: promises and challenges. IEEE Computer 33, 7, 36–41.
KOSHIMA, H. AND HOSHEN, J. 2000. Personal locator services emerge. IEEE Spectrum 37, 2, 41–48.
MOBILE MARKETING ASSOCIATION (www.mmaglobal.com).
NORTEL’S MOBILE LOCATION SERVICES SOLUTION. 2002. (www.nortelnetworks.com/corporate/news/
newsreleases/2002c/07 02 02 wollongong mlc.html).
PAYBOX (www.paybox.net).
RASTIMOR, O., KOROLEV, V., JOSHI, A., AND FININ T. 2001. Agents2Go: An infrastructure for location-
dependent service discovery in the mobile electronic commerce environment. In Proceedings of
the 1st ACM International Workshop on Mobile Commerce (Rome, Italy, July). ACM, New York,
31–37
SASTRY, D., COLE, A., MUNSON, J., AND CHRISTENSEN, J. 2001. An approach to providing a seamless
end-user experience for location-aware applications. In Proceedings of the 1st ACM International
Workshop on Mobile Commerce (Rome, Italy, July). ACM, New York, 20–25.
SPRINT AND eONE MOBILE PAYMENT NETWORK. 2002. (www.mcommercetimes.com/Services/254).
TRACZ R. AND WRONA, K. 2001. Fair electronic cash withdrawal and change return for wireless
networks. In Proceedings of the 1st ACM International Workshop on Mobile Commerce (Rome,
Italy, July). ACM, New York, 14–19.
VARSHNEY, U. 2001. Addressing location management issues in mobile commerce, In Proceedings
of IEEE Conference on Local Computer Networks (LCN-26) (Tampa, Fla., Nov.). 184–192.
VARSHNEY, U. 2002. Mobile commerce: Issues, technologies, and solutions, A tutorial at ACM
International Conference on Mobile Computing and Networking (Mobicom) (Atlanta, Ga., Sept.)
ACM, New York.
VARSHNEY, U. AND MALLOY, A. 2001. An integrated approach for improving the dependability of the
emerging wireless networks. In Proceedings of IEEE International Conference on Global Com-
munications (Globecom) (San Antonio, Tex., Nov.) IEEE Computer Society Press, Los Alamitos,
Calif., 3693–3697.
VARSHNEY, U. AND VETTER, R. 2002. Mobile commerce: Applications, frameworks, and networking
support. ACM/Kluwer J. Mob. Netw. Appl. (MONET) 7, 3, 185–198.
VARSHNEY, U., VETTER, R., AND KALAKOTA, R. 2000. Mobile commerce: A new frontier. IEEE Com-
puter 33, 10, 32–38.
VODAFONE MOBILE PAYMENT SYSTEM. 2002. (www.pcmag.co.uk/News/1128273).

Received June 2002; revised July 2002, August 2002, January 2003; accepted February 2003

ACM Transactions on Internet Technology, Vol. 3, No. 3, August 2003.