GPRS: At last the wait is over

Viag Interkom launched GPRS (General Packet Radio Services) handsets to the German public last week. It hasn't said how many handsets are available. It probably isn't a large amount. But it's the first consumer GPRS launch in a major European market.

In the UK, Viag's sister network BT Cellnet was expected to launch GPRS to the general public before the end of 2000. It hasn't happened, and Stuart Newstead, general manager wireless data services, told silicon.com that it wouldn't happen until "late spring or early summer 2001".

So GPRS has only just dipped its toe into the consumer market. It'll be a long time before it takes the plunge completely. What's more, the technology that a few lucky Germans are playing with today isn't even the GPRS tech watchers have been dreaming about.

For one thing, it's slower than expected. Nine months ago, research reports and industry white papers were quoting speeds of over 100Kbps. Viag and Deutsche Telekom's GPRS services will be shipping with 20Kbps speed, and the possibility of 40Kbps at a later date.

Of course, it's not all about speed. The joy of GPRS isn't pumping massive quantities of data back and forth, but providing an 'always-on' connection. That's what users of Japan's iMode mobile internet service have been enjoying, and it's one of the main reasons why iMode has been such a success.

Why is GPRS late? One possibility - rarely mentioned - is that there is a health risk. Higher data rates require more energy, and more energy means more potentially harmful radiation. If second-generation (2G) GSM phones are still causing health scares, how big will the risk be from faster, 2.5G GPRS phones?

This speculation may be far-fetched. A DTI spokesman admitted it is conceivable GPRS devices, "designed or used inappropriately, could produce exposures above the levels specified [by the European bodies determining safe levels]," but they haven't seen any GPRS phones that are dangerous in normal use.

It seems far more likely, in fact, that the delay of GPRS is down to technical, not health factors. Whatever the health implications may be, the major difficulty with providing faster GPRS services is battery power.

The current GPRS phones - or more specifically, phone, because only Motorola's T260i is currently available - uses two 'slots' on the uplink to deliver more data. A normal phone only uses one slot.

Rates are currently expected to be around 20Kbps. That requires twice as much energy as a single slot link, which a normal GSM handset uses. Finding batteries to power all this data is extremely difficult - especially as they have to power enhanced display capabilities as well.

When two more slots are added to make a four slot link, the power factor doubles again. It's not only batteries that become an issue here, but heat dissipation too. The phone is at risk of overheating, and there's no way to build in a giant, PC-style heat sink and fan.

To get the full 100Kbps+ data rates promised in early GPRS white papers, eight time slots are needed. "There's no phone in the world that could run all eight time slots. It would just get too hot," said Nigel Deighton, a vice president at research company Gartner.

A third possible stumbling block is the network. GPRS requires extensive modifications, with either a software upgrade of existing base stations or the replacement of older ones.

Viag Interkom, for example, is a relatively new entrant to the market, without the mix of network infrastructure that a larger, older player like Deutsche Telekom may have. "People who rolled their core networks out later may have newer infrastructure," said Deighton. "That might enable them to roll out GPRS faster."

An even more vexing question is the matter of billing. Billing for circuit-switched data is easy - a call is charged by the amount of time a customer spends online. It's easy to understand, simple to implement and there is no need to track what a customer does.

Billing packet-switched-based services such as GPRS is much harder. Not only do billing systems have to be updated, but whole new business models have to be devised - and then communicated to users.

BT Cellnet clearly has had some trouble working out a billing arrangement for its business clients, so how much harder will it be to devise a model for consumers, already confused and sceptical after the WAP fiasco?

BT Cellnet's Newstead confirmed the company will be "prioritising business customers". That's not surprising. Consumers are less tolerant of faults in new hardware, and less likely to pay large sums of money for unproven or confusing services.

Gartner's Deighton added: "I think that's the right approach for BT, at least at first. I think for the next one or two years the main customers will be business people. In the next three or four years consumers will have caught up, but for now the business users are the right place for operators to be concentrating."

It's an argument that makes sense. However, NTT DoCoMo's experience with its proprietary iMode in Japan - where a GPRS-like mobile data service has been a massive hit with consumers - suggests there is money to be made from Joe and Jane Public.

Launching services to consumers too early, before there are lots of handsets and bugs have been ironed out, would be a mistake. But with massive 3G licence fees and infrastructure investments to pay for, BT Cellnet and other mobile operators need every source of revenue they can get.

And any delay in developing consumer applications for GPRS will hamper operators' access to revenue streams from the next stage in the evolution, namely full blown 3G.