WiFi comes to rescue of GSM in rural broadband test

by Guy Kewney | posted on 27 November 2002

It may seem unlikely - after all, most commentators now believe that WiFi networks will destroy the market for mobile phones - but the boot could be on the other foot, and 802.11 technology might even extend the reach of GSM phones. Invisible Networks and ip.access are to trial a hybrid system.

Guy Kewney

You can build a new 7-slot GSM base station for "well under five thousand pounds," according to Stephen Mallinson, Managing Director of ip.access - if you use low-cost, mass-production chips and feed the station over the Internet.

And if you haven't got a low-cost Internet, you can get one, by setting up long-distance WiFi nodes. This is now going to be tested to see how well it works, in a trial - the first of its kind anywhere in the world - to the north of Cambridge.

The GSM base stations will piggy-back on the Fens Community Broadband Project, which was set up by Invisible Networks, and will not be part of any of the big mobile operators. Instead, it will operate as a private network with a test and development licence. Invisible and ip.access "will be inviting mobile network operators to view the solution and see for themselves how their rural coverage problems can be easily and cost-effectively solved."

The plan is a cross-subsidy one. On one hand, you have rural communities with no cable, no ADSL, and leased lines costing an arm and a leg, wanting broadband. And on the other hand, you have GSM operators finding that there are large geographical areas where they have no masts - and where the locals oppose planning permission for new ones.

<1/> Stephen Mallinson

"Our technology can be slung on the back of any IP based network," said Stephen Mallinson, managing director of ip.access. "It's a picocell box, a very small base station, which has an IP connection. It becomes a very simple thing for the operator to set up. All they need is the IP address of the base station."

Because the picocell is so small and low-powered, it can be set up without planning permission, and without alarming neighbours. "It's a thing about the size of a small Smartie tube, and you clip it to the outside wall of a house," said Mallinson. One picocell can support a village of 100 cellphone owners with ease, based on normal traffic per phone.

What hasn't been proved, yet - and what the trial is supposed to prove - is that you can provide GSM phone links over an IP connection, without introducing latency. There's a maximum delay which the GSM specification will allow between sending a signal, and hearing the voice at the other end.

"Our own internal trials show that this is an issue, but that it isn't a problem. There are problems which we've had to address with jitter; but latency hasn't been a problem," said Mallinson. "We have an internal test linking us to Vancouver, where the latency is acceptable. However, we aren't taking chances, because in this test, we're sharing a broadband backbone off Invisible Networks, and so the wiring is physically separated out, in the wiring closet, so that the one that carries GSM is separate from the one that takes normal LAN traffic. This makes it more secure, and protects you from latency caused by heavy LAN users. The sockets are powered intelligently, too; so while the picocell TRX gets its power, normal LAN outlets don't carry power which might damage equipment.

The synergy could break a log-jam. Invisible needs a critical mass of would-be broadband subscribers before it can justify installing one of its rural wireless-based broadband services. Large mobile carriers need to find a way of serving remote areas, without the huge expense of putting up full size masts. And ip.access has a sub-£5,000 GSM cell which is adequately-margined, but vastly less costly to install and operate than any alternative on a leased line.

What will happen now is:-

1) Invisible will target a group of villages where there is some interest in broadband.

2) Even if the initial interest is too small to justify it, they can subsidise the installation costs by getting one of the big carriers to sign up

3) when this is sorted, they install a leased 2 megabit line, connect a wireless 802.11b transceiver in one site, and several subsidiary transceivers on house tops. The GSM picocell goes live straight away, and starts generating revenue for both Invisible and T-Vodagenange, or whoever.

4) Local residents see the system working, and start to become interested.

A good example of a rural broadband community is the Six Villages project.

More about ip.access from their site. More about this story here, later ...

And ip.access ltd "designs and manufactures a range of GSM and GPRS infrastructure solutions that integrate seamlessly with IP-based networks providing ubiquitous and cost-effective mobile radio coverage within buildings. ip.access products can be used to enhance public mobile networks and to support emerging enterprise-focused wireless office services. Based in Cambridge, UK, ip.access ltd is a wholly-owned subsidiary of TTP Communications plc, a world leading independent supplier of cellular handset technology. TTP Communications, through its subsidiary TTPCom (, sells technology intellectual property used in the design and manufacture of wireless communication terminals. Over 60 devices using its GSM technology have achieved Full Type Approval - more than any other independent developer. TTP Communications plc is listed on the London Stock Exchange (LSE: TTC)."

And Invisible Networks "is an East of England based private limited company, formed to deliver Community Broadband to areas that the big telcos are avoiding. The company brings together Internet Networking expertise and Wireless Networking knowledge to deliver broadband services to rural areas. The combination of a flexible low cost model and the frustration felt by users bypassed for Broadband has meant very fast growth for the company that only launched the first live networks in September 2002."