Mobile's missing links

Mobiles that can pull content off the Web in moments and run video-based killer applications? Don't you believe it, says Sanjay...

Mobiles that can pull content off the Web in moments and run video-based killer applications? Don't you believe it, says Sanjay Manandhar, head of e-business integrator Rubus

By just about any measure, WAP was a hype that failed to deliver. WAP (Wireless Application Protocol) was a classic case of 'overpromise and underdeliver', which is a dangerous cocktail for a nascent market. Now it looks as though the next-generation mobile infrastructure standards, GPRS (General Packet Radio Services) and UMTS (Universal Mobile Telecommunications System), are likewise raising users' hopes preparatory to dashing them.

Hardware 1, software 0

The migration path for mobile infrastructure is surprisingly well understood by the general population. People know that the current mobile infrastructure is in its second generation (2G) and is digital as opposed to the analog first generation. Now, we're waiting for 2.5G devices using standards such as HSCSD (High-Speed Circuit-Switched Data) and GPRS, as well as 3G phones, which will support Edge (Enhanced Data for GSM Evolution) and UMTS. You might conclude that such awareness of arcane technologies is evidence of the power of telecoms marketing, and that WAP is little more than an attempt by mobile operators to find a mechanism for generating revenue from users wanting to access the vast content created on the Internet.

The table shows some of the key attributes of various network infrastructures. Both bandwidth and timing are theoretical and aspirational. For instance, in HSCSD trials to transfer data at much faster speeds using four GSM radio channels simultaneously (four times 14,400bps or 56.7Kbps), the best we've been able to get from a local mobile operator is two times 9,600bps or 28.8Kbps [CHECK].

Despite a clear migration path for mobile infrastructure, what is clearly missing is the middleware and applications that use the higher bandwidth effectively and deliver services that users are willing to pay for. For GPRS, Edge and UMTS to be successful, the applications and services they underpin must be compelling enough for users to migrate to faster data speeds. High-quality video phones or video streaming of events are some of the bandwidth-hungry applications that have been thought of. It isn't clear whether these will be killer apps; the only way to find out is to release the protocols and technologies into the public domain and create a competitive industry around the infrastructure.

Packet problems

GPRS differs from the current network infrastructure in that the data is packetised, which makes better use of the radio spectrum. Despite a new network roll-out, GPRS, which is touted as the network infrastructure that will solve all the difficulties that beset WAP, could be sorely misplaced.

It is hard to believe that GPRS as a service will be attractive by the end of this year. To realise why this is the case it is only necessary to consider each of GPRS's attributes in turn in the light of realistic rather than hyped-up parameters.

  • Packet-switched. Since GPRS depends on packet-switched rather than circuit-switched data, it makes very efficient use of valuable spectrum. The mobile operators would love all subscribers to move to packet-switched networks. However, making the network packet-switched means a huge network software upgrade costing something like £150m for a UK-wide national network upgrade.

  • Base stations. Currently, each operator has a few thousand base stations for GSM coverage. Rolling out GPRS data will mean increasing the base station count to about 10,000. In addition to the physical deployment of hardware, the right-of-way issues could seriously hamper radio coverage and quality of service.

  • End-user devices. The current generation of handsets cannot be used for GPRS, so new handsets will be needed. But with UMTS requiring modulation changes, GPRS handsets will be useful for only a few years at most before being forced to retire.

  • Pricing. Although current pricing schemes subsume most of the handset costs in airtime, the pricing mechanism could deter widespread or quick penetration. Many operators have a number of pricing models under review but aren't sure which to use as ease of understanding is key to a popular service.

  • Always on. This feature of GPRS has received a lot of popular support, especially as it would solve many of the problems of WAP applications. However, the GPRS network only simulates a continuous session, so when there is a break in the connection a second or two of setup time is required to re-establish the session. Hence, while the call setup time (as with WAP) will be significantly reduced, the GPRS connection is not always on.

  • Bandwidth. The first devices to be released will have bandwidth capabilities closer to today's GSM speeds. By the end of this year, systems with data speeds of 43.2Kbps to receive and 14.4Kbps to transmit should be possible. By the middle of 2001, bi-directional data speeds of 56Kpbs should be possible with bursts speeds of up to 115Kpbs.

    The theoretical limit of the network is 171Kpbs, which will probably never be achieved.

    So will GPRS be ready for general public use by the end of this year? I doubt it and would not expect reasonable penetration until the end of 2001.

    Probably the greatest contribution of GPRS will be better usage of precious spectrum, as we move from circuit-switched (dedicated logical connections) to packet -switched connections. In addition, the packet nature of the traffic means billing models can't be 'per minute' alone because nobody 'owns' a physical circuit. Innovative billing models will have to be tested and explained to users who have been used to per minute voice charges.

    What About 3G?

    Although there is much hype around 3G, fanned by the theatrical auctioning off of spectrum licences, the cost-benefit analysis is far from complete as the applications and likely takeup is not fully known. There will clearly be user segmentation, and early adopters and business users may pay a premium for higher bandwidth, which may not justify wholescale network upgrade.

    While Edge (an interim network upgrade from GPRS to UMTS) can provide data speeds of up to 384Kbps, bandwidth capability does not imply success in the market. UMTS requires modulation changes from the current TDMA (Time Division Multiple Access) to CDMA(Code-Division Multiple Access), which will be incorporated into Edge, which is to be deployed in 2002. Edge will require upgrading the base station hardware, putting the capital cost of national rollout in a country like the UK at £300-500m.

    UMTS or 3G is an international effort to co-ordinate spectrum allocation and modulation and other technical specifications so that global roaming with a single handset would be possible. Although 3G data speeds can reach 2Mbps, 384Kbps is the more likely scenario in 2003. As with Edge, middleware and access to network simulation/trials need to be forthcoming if the network is going to be useful for people by that time.

    If the mobile operators, device manufacturers and infrastructure suppliers continue to promote 2.5 and 3G networks, we will receive just the networks and no real services riding on them. What we need are middleware and revenue-generating applications that come to the market in tandem with the roll-out of the networks.

    Mobile Infrastructures

    Generation 2G 2.5G 2.5G 3G 3G
    Bandwidth 9.6Kbps 57Kbps 171Kbps 384Kpbs 2000Kbps
    Payload Text/voice Graphics/files Graphics/files Video Video/multimedia
    Timing Now Now End 2000 2002 2003
    Network type Circuit Circuitx2/4 Packet Packet Packet
  • Read more on Wireless networking