News

Satellite: that's the answer to keeping train WiFi costs down!

Internet for passengers "still provides a compelling reason for increased use of trains," said 21Net chairman Henry Hyde-Thomson today. The solution, he suggested, is to keep capital costs down - and that, he insists, means satellite connectivity.

His company was a prime mover in the recent announcement of Thalysnet,  which was announced back last September, and went into use on the Eurostar last month.

"New gadgets like the iPhone will drive usage; other WiFi phones and hand-held internet devices will become more and more popular, becoming the new standard… so internet access demand isn't just going to come from business travellers with laptops," Hyde-Thomson said.

His 21Net service started off running a prototype 4 megabit services in 2005, but today, has stepped it up to 40 megabits (downlink) with a 2 megabit uplink system in modern satellite links. "We reckon a minimum of 2.5 megabits is needed for a train," he told delegates at the Traincomms 08 seminar today.

Cost are based on what the satellite operators charge. "A whole transponder costs between 1m and 5m Euros per year, giving 40 MHz; if you look at the Mbps ratio to MHz, which varies, you can improve the costs."

The problem, he said, is that you have a finite amount of power on the satellite. Increase the data error correction rate, and you cut the throughput; increase the bit rate and you become more vulnerable to obstructions to the signal. Use a large dish, and you get better performance - typically, a high gain antenna will give up to ten times the megabits per MHz.

So, for a fleet of 25 trains, a low-gain satellite dish on the train is very expensive, especially if you don't use spread spectrum, giving 250 MHz for a 2 megabit link. But if you optimise with a high gain antenna and uses DVBS-RCS satellite tech, you can get the same bandwidth out of only 11 MHz of satellite.

Typically, the capital expenditure on the satellite is lower than the other trackside infrastructure (such as Wi-Fi at stations).

Ruefully, Hyde-Thomson admitted that the case for broadband for passengers in the UK is not quite a compelling.

"The UK experience is not representative of the whole of Europe, and clearly, the advantage of broadband on trains goes away if everybody is jammed together!" he said. "So we have been more focused on Europe; they are not overcrowded; normally, you have a table, or a flatplan."

Should train operators consider running two networks - one for passengers, and the other for train management? He said probably not: "If you combine the case for combining business service to passengers, you have a better business case."

In the case of Thalys, they currently have handheld PDAs for train staff, and these use GPRS, and these will migrate to onboard WiFi in time; there may be other apps in time as well, he suggested. These won't reuiqre completely separate networks: "Unlike some, I think there is a good business case to providing broadband to passengers ."