Despite the fact that the Global System for Mobile Communications (GSM) network was designed from the ground up for data as well as voice, its use for wireless data transmission has been slow to catch on. In 1997, only 2 percent of GSM network revenues were generated by data communications.
However, the Short Message Service
(SMS), a pager-like data service that transports brief text messages, has been one of the most successful enhancements of the GSM network. SMS enables users to receive dispatches of 160 characters and to read regular e-mail on a mobile phone. It has become the fastest growing GSM service. That's why the hopes of network operators and equipment vendors are high for what is collectively called
smart messaging. This new technology could soon bring the Web paradigm of standardized interactive communications to mobile data users.
With the opening of the European telecommunications market this year, GSM network operators all across the continent are now looking for new ways to differentiate their businesses. What they want are faster, cheaper, smarter, and better-managed data services to expand the market. "Market expansion is only feasible if we can offer the mobile workforce better integration with corporate information technology," says Bob Apollo, managing directo
r of XcelleNet International, a multinational GSM solution provider.
A good example of new data services is Vodafone's Data Direct in the U.K. It manages all of an enterprise's mobile messaging and other wireless data traffic. For example, Vodafone handles British Gas Services' integrated voice and mobile data system that is used by a field force of 5200 engineers. It includes a vehicle positioning system that works without the need for expensive Global Positioning System (GPS) equipment. A project coordinator can determine the location of field operatives via the regular network poll of the GSM phone and at the same time distribute critical project information.
Although several of these new wireless enterprise data services are emerging, the wide use of GSM for data hasn't really caught on. What are the reasons for this sluggish growth? There are basically three key issues.
Low transfer rate of 9.6 Kbps limits the scope of mobile data applications.
However, through aggregation of time slots on one channel, operators will eventually be available to provide higher bandwidth on demand.
Although mobile-phone vendors and network operators, as well as standards bodies, are addressing all these issues, the reference point for the quality of service is the fixed network. Says Mike Dreisch of Intercai, a consultancy in the Netherlands, "To capitalize on the growth of corporate Internet and intranet usage, the mobile network will need broadband links comparable to the fixed services within two years. Otherwise it
will have a problem." Martin Garner, a market analyst at Ovum's research group, notes, "Cellular operators need to watch the fixed network very carefully. They will have to carry on improving call-setup times, middleware, compression, and protocols in order to keep up."
Wireless Application Protocol
One of the key concepts of the next-generation data-centric GSM network is the Wireless Application Protocol (WAP). It is designed to enable delivery of interactive information to mobile devices (see "Smart Messaging to Come to GSM," January BYTE international edition). "We see WAP as an excellent opportunity to provide our customers with access to a wide range of services and information," says Ian Germer, marketing executive at service provider Vodafone.
Various vendors such as Nokia, Motorola, Ericsson, and Unwired Planet have been offering different versions of this smart messaging concept, but an industry consortium called the WAP Forum now administers the specification process. Initially,
WAP was targeted at GSM networks to bring standards-based applications and Internet content to cellular phones. But because the WAP specifications shield applications from the underlying network, it can easily be deployed on other wireless networks as well (see the figure
"WAP Architecture"
).
WAP defines a set of protocols for the transport, session, and security layers. This concept allows the transport layer to adapt to specific features of the network, yet at the same time it achieves consistent global interoperability using mediating gateways. The session layer defines a common data exchange mechanism to which more specialized session protocols can be added. External applications can also access the session and transport layers directly, but direct access from external applications to the security layer is not foreseen.
Scalability, both on the device and in the network, is another important issue that WAP addresses. The framework and the applications can be employed on regul
ar one-line-display phones as well as on advanced PDAs and hand-held PCs. WAP applications can run across bearer networks with different bandwidths. The architecture is also designed to accommodate future bearers such as
General Packet Radio Service
(GPRS), which is expected to appear within the next two years.
The key architectural components include a Wireless Markup Language (WML) browser, WMLScript interpreter, Telephony Value-Added Service (TeleVAS) framework, and the three protocol layers. WMLScript is designed to reduce overall network traffic and thereby make optimum use of GSM's 9.6-Kbps air interface and the relatively limited resources of the telephone. Like Java, this technology enables applets to be transmitted to the client device. In other words, WMLScript helps download intelligence into a mobile device.
TeleVAS is a concept that network operators can use to enhance or build services. It provides a device-independent interface to call-control, phone-book, and mes
saging functions. Via TeleVAS, users can download the menus of the services to which they subscribe, thereby having an easy method of accessing the value-added services. Existing services that use touch-tones, for example, can employ a TeleVAS application wrapper that presents the user with a scrollable menu. This will let operators seamlessly integrate Web and telephony applications in existing solutions. Additionally, smart applications can persist in local memory for quick access, and updates can be downloaded to the phone handset whenever network services are added or dropped.
WAP or Java?
The programming model of WAP closely follows that of the Web. All content is specified in formats that are similar to those of the Internet. It builds on the concept of content interpreters, location-independent addressing, and event handling. Plus, WAP allows all content and applications to be hosted on standard Web servers using technologies such as CGI and Java.
WAP aims not only at value-added ser
vices for mobile users. It could also be used to provide mobile workers with secure, wireless access to corporate information over the Internet. This concept is often referred to as the "mobile intranet."
WAP and Web technologies such as Java, at least for the next few years, will be complementary and not competing technologies. Smart messaging and WAP are meant for "lite" GSM phones, whereas Java aims at more sophisticated network terminals such as Nortel's Network Phone (see "What's Hot at CeBIT," April BYTE international edition). Adding WAP functionality does not require any significant increase in handset resources, so the price of WAP-enabled phones would stay about the same. The mainstream market is unlikely to go for any technology that increases costs.
Although phone display sizes and computing power are increasing, they are still a limiting factor when it comes to running complicated Java applications. WAP offers the option to strip out all graphics from a standard Web page, so it better fit
s the needs of wireless networks. On the other hand, more general-purpose devices that aim at Web browsing might employ Java as the main application implementation environment. For example, WAP could be used on a standard phone to deliver traffic and weather information, while Java could provide stock-exchange data graphically on a hybrid PDA/phone.
The crucial question is whether the WAP concept will reach critical mass, considering that Java is already a pretty solid standard. According to Nortel's director of business development, Ken Blakeslee, "the thirst for interactive services will not end with smart messaging but will leverage the critical mass of Java developers and applications."
More Bandwidth, More Service
More intelligent WAP- and Java-enabled devices will help a company better support an outside sales force, manage fleets, or control deliveries. But what will throughput service providers offer customers?
Later this year, GSM operators will be able to offer a 14.4-Kbps ser
vice that employs a slightly less robust air interface coding. And around 1999, multiple channels will be available on a "demand basis" as so-called High-Speed Circuit-Switched Data (HSCSD) services. "With the introduction of HSCSD, a whole new range of wireless applications, such as mobile videotelephony, will become viable," notes Pekka Pohjakallio, head of Nokia's product marketing.
HSCSD uses time division multiple access (TDMA) technology to divide each channel into several time slots. The theoretical maximum transmission rate is eight times 14.4 Kbps, which yields a throughput of 115 Kbps. The higher 14.4- Kbps data rate in a single slot comes simply as a result of decreasing the amount of redundancy in the transmission channel.
Several operators have indicated that they are planning to start with a two-slot 28.8-Kbps HSCSD service and then offer an asymmetric combination of four slots for Internet access or videoconferencing. This will provide a downstream rate of 43.2 Kbps and an upstream rate
of 14.4 Kbps.
General Packet Radio
As its name implies, HSCSD is a connection-oriented protocol. It is therefore ideal for real-time applications that require low latency, such as videoconferencing. However, future GSM networks will also provide a packet-based service. In many respects, for operators, the move to the so-called General Packet Radio Service (GPRS) is a more significant change than the introduction of HSCSD.
GPRS is basically a packet-switched overlay network with four different coding schemes and four different single-channel data rates, from 9.05 to 21.4 Kbps per channel. Up to eight channels can be used, providing a maximum data rate of 171.2 Kbps. It is much better suited to data applications such as Web browsing, e-mail, and database queries. Packet switching also allows more users per network, so the economics are likely to be more attractive for the users.
As with HSCSD, changes to the handsets are necessary to accommodate packet switching, and additional investmen
ts have to be made to leverage the infrastructure. But interconnection to other packet-switched networks, such as the Internet, will help operators provide flexible data services for enterprises, which could pay back in a couple of years. The biggest change from an operator's perspective will come from the need to adapt the business model from one based on connection time to one that charges on a "per megabyte" basis.
Wireless Internet Access
Today most cellular-network operators are missing out on the business opportunities that flow from robust Internet connectivity. The main reason is that remote access from a GSM client currently requires either a telephone or an ISDN connection from the operator's mobile switching center to the corporate LAN. This requires a corporation to deal with a pool of modems or ISDN boards in the remote access server and to pay for the fixed access lines to the LAN.
GPRS will change this scenario. With GPRS in place, cellular operators will be able to source d
ata traffic directly into the corporate LAN via the Internet and a standard router. GSM operators will then be able to work as wireless IP service providers. As a result, the cellular network can operate as just another segment of the corporate intranet.
Commercial GPRS services are unlikely to appear before the year 2000. GPRS may be the superior technology for most wireless applications, but for operators, the move to packet switching and the corresponding business model is not a trivial matter. Not all network operators will be able to justify investments in all GPRS upgrades in the near future. Because the transition to HSCSD is much easier to make, some operators will therefore stay with HSCSD.
However, if operators in different regions provide different environments for mobile data, roaming will be possible using only the regular 14.4-Kbps service. In addition, some operators may even hold back and wait for the third generation of services, dubbed Universal Mobile Telecommunications System, or
UMTS (see sidebar "Getting Closer to a World Mobile Phone"). This technology is expected to make an appearance around 2002. One of the key objectives of UMTS is the definition of a unified worldwide wireless broadband infrastructure.
The Future of GSM
The future of mobile data services will be determined by the volume of traffic generated in the next two years, as well as by the progress of UMTS. If traffic grows, GPRS will be widely implemented and eventually be complemented by UMTS.
Once the GSM infrastructure has been upgraded to handle packet-switched data, and the medium has proved its effectiveness, then network operators may be able to enhance their offerings with Internet telephony. By that time, low-latency, premium-rate Internet services based on IPv6 will go all the way from the Internet backbone to a standard GSM mobile client. The mobile phone will then be able to deliver a new level of convenience to corporations and consumers alike.
illustration_link (43 Kbytes)
WAP shields applications from the underlying network but supports network details through gateways.
illustration_link (35 Kbytes)
With General Packet Radio Service, the corporate LAN obtains traffic from the GPRS network.
Bob Em
merson is a freelance writer in Eindhoven, The Netherlands. You can contact him by sending e-mail to
bobe@IAEhv.nl
.
The Mobile Intranet
