 |
|
|
The second generation of cellular systems (TDMA and CDMA) is still in the early stages of installation around the world and the first generation is still growing (AMPS and other analog systems). Yet everyone, apart from the birds and the bees, is now talking about the third generation.
The interest in the third generation of wireless has been focused by an ITU initiative known as IMT-2000 (previously known as FPLMTS). This provides requirements to be met by any technology to ensure that multi-media can be supported along with traditional voice services. The main requirement for multi-media (data and video) is high data rates, from 144 kbps for high mobility applications (e.g. phones in vehicles) to 2 Mbps for stationary applications.
There seems to be almost universal agreement that CDMA technology will form the third generation radio interface. Even Ericsson, the staunchest TDMA proponent, is now promoting its W-CDMA concept. This new generation radio interface is known as Wideband CDMA because it uses a 5 MHz radio channel, as compared to the 1.25 MHz channel specified by today’s IS-95 based CDMA systems.
It is worth pausing to reflect on the almost universal promotion of CDMA as a third generation technology. After emerging from years of debate over FDMA versus TDMA versus CDMA, a consensus has emerged. The US, for example, has commercial systems using FDMA (NAMPS), TDMA (IS-54, IS-136 and GSM) and CDMA (IS-95). It seems incredible that the industry may, in the future, reduce the number of incompatible technologies. Optimism must be tempered, however, as fragmented standardization could result in multiple incompatible CDMA standards.
Spectrum availability is always a stumbling block for any radio technology that aims for global application. According to a Motorola report (“Third Generation Systems Briefing”), while the ITU recommendation for the use of spectrum in the 2 GHz area has been adopted by Europe, Africa, the Middle East, Japan and Korea ... the United States and Canada have allocated much of this spectrum for second generation PCS, and Latin America is following suit. Other countries are expected to follow a mixture of these two approaches. Consequently it is almost certain that a universal third generation standard would have to support at least two different spectrum bands.
An important, but often overlooked, technology is the network, which provides roaming services such as call delivery, inter-system handoff and service mobility. There are two choices: the GSM MAP (Mobile Application Part) and ANSI-41 (formerly IS-41). GSM MAP is currently only used for systems using the GSM radio interface, but it could be adapted to support other air interfaces. The TIA standard ANSI-41 (or, more formally; ANSI/TIA/EIA-41) currently supports several different radio interfaces – analog, NAMPS (IS-88/IS-91), D-AMPS (IS-54 and IS-136 TDMA systems) and CDMA (IS-95). The Ericsson/Nokia W-CDMA approach claims to be based on the GSM MAP, while a system being promoted by Lucent, Motorola, Nortel and Qualcomm is based on ANSI-41.
NTT DoCoMo is promoting the use of INAP (Intelligent Network Application Part) for third generation systems. However, as this protocol has never been used for mobility applications it is most likely to find use as an adjunct protocol for providing advanced services, rather than as the backbone of a mobility network.
There is a recognition among key players that compatibility must be provided with older systems. Without a migration path, it is unlikely that entrenched carriers can be persuaded to invest in a new technology. Multi-frequency, multi-technology phones will also be a factor for some time, as consumers are unlikely to purchase phones that do not work as far afield as existing phones. Most people would be willing to sacrifice high speed data capabilities outside their home service area, as long as their phone can continue to make voice calls.
Another aspect of compatibility derives from the two existing network standards (GSM MAP and ANSI-41). These standards are so different that interworking between them is not strictly possible. One example of this is authentication, which uses different data formats, different strategies and different algorithms in the two systems. Ericsson promotes the concept of an ILR (Interworking Location Register), basically an HLR with both an ANSI-41 and GSM-MAP interface. In order to fully use this concept, phones will have to be designed with support for both networks.
It might seem that network protocols should be relatively independent of the radio access protocol. This is true to a considerable extent for the lower layers, but not for the upper layers that control services and capabilities that are managed by elements deep within the network. For example, GSM mobiles control features such as call forwarding by transmitting special radio interface messages that identify the specific feature. Mobiles on ANSI-41 networks simply transmit a string of digits that are interpreted by the mobile's HLR. One can argue over the advantages of each method, but the important point is that the network signaling for the two methods is not likely to be the same. Similar comments apply to authentication, handoff and other capabilities.
Carriers have invested extensively in wireless technology to provide first and second generation service in many countries. They will have to be persuaded that their investment will not be lost if they provide third generation service. The press releases on W-CDMA and other IMT-2000 compliant systems are trying to define this vision. The purpose of trials, such as that scheduled to start in 1998 by NTT DoCoMo, which will utilize Ericsson, Lucent, Nokia and Siemens equipment, is to establish the feasibility of the vision. Much is at stake, because companies that do not successfully project a vision of the third generation will likely be excluded from many second generation purchasing decisions.
IMT-2000 promotes the concept of mixing voice with data on wireless systems. Will higher speeds for data finally make it as important as voice on wireless systems? I am highly skeptical of such claims, and simple mathematics will show why. Voice coders for third generation systems will likely be 8 or 16 kbps, possibly as high as 32 or 64 kbps. With 8-16 kbps speeds, it should be possible to achieve toll quality voice, especially considering that wireline systems use 32 or 64 kbps voice coding schemes, without the sophistication of more modern systems based on complex algorithms executed by high speed DSP’s. The voice coder bit-rate will likely go down over time, as voice coder technology improves.
Data rates, on the other hand, will continue to climb as wireline systems continue to advance and as applications make more intensive use of graphics, sound and video. Wireline internet surfing will continue to raise the expectations of users. In the last several years, data rates in mass market modems have climbed from 9.6 kbps to 14.4 kbps, 28.8 kbps, 33 kbps and now 56 kbps. By the time IMT-2000 compliant systems are in the field (early in the next millennium), the data rates that internet surfers have grown used to will be considerably higher (already xDSL systems are available that run over standard copper wiring at multi- Mbps rates). Assuming a constant or declining bit rate for voice coders, this will mean that the number of equivalent voice channels that a single data user occupies will climb higher over time. Since the cost of airtime for a voice call is unlikely to drop dramatically, the cost of data must rise over time.
The management of radio resources to allow voice and data to coexist will also become more complex. It will be like a parking lot user trying to make room for everything from bicycles to Mack trucks. One 384 kbps data user parking on a video conference call will displace 24 potential voice users (using 16 kbps coders). Will the video conference user be prepared to pay 24 times as much? Or will carriers be willing to reduce their charges for voice to make data cost effective? Another important issue is the risk of congestion caused by even a small surge of data users.
There may be a single standard at the end of the search for a third generation wireless system, although that is far from guaranteed. What is certain is that standards organizations will fight for pieces of the action. The TIA and ATIS will be active in North America, supported heavily by US carriers and vendors. ETSI will be equally active in Europe, supported by their carriers and vendors. Whether they can set aside their differences, and use the ITU to work jointly is a critical question. If not, ATIS and ETSI will probably provide a standard that looks much like a beefed-up GSM and the TIA will provide a system that looks like IS-95/IS-41 on steroids. It would be sad if this happened, because the likelihood of seamless international roaming will be diminished.
IMT-2000 provides a real opportunity for international cooperation between carriers and vendors. As the demand for international roaming increases, the demand for one phone that works everywhere will also increase. It is almost certain that the base technology of this phone will be 5, 10 or 20 MHz CDMA. If it is to work both in traditional GSM regions and traditional AMPS regions, it will have to support both network protocols. A risk is that the third generation technology will try to provide too many disparate services and become bogged down by complexity and high costs.
© – Copyright