News

Voice over Wireless IP - the Achilles Heel of 3G?

Analysts are waking up to the potential threat posed to expensive 3G phone networks by wireless LAN technology - but appear to be (wrongly) convinced that the two systems can't compete.

The disruptive potential of voice over IP (VoIP) appears to have eluded several observers. Quoting several industry analysts in Business 2.0 reporter Matthew Maier suggests: "Although both WiFi and 3G provide high-speed wireless access, the two systems are actually very different."

Third-generation networks offer more efficient voice and data transmission -- with voice calls remaining the most popular application, he summarises. "In contrast, WiFi systems are used only to transmit data."

In fact, voice is data. Data-only LAN traffic already includes a vast amount of digitised voice traffic, as offices connect their PABX technology to Ethernet, and route internal calls to branch offices over the data links, including cheap Internet links.

And people are making money out of this. Companies like Net2Phone and CallServe, startups dating from the dotcom boom, are still growing - offering computer users cheap voice calls over Internet, direct to dialled landline phones.

The one thing that VoIP still can't do, is receive phone calls independently of the established telcos, who have a monopoly on international dialed phone numbers. So while I can use CallServe to dial your phone number over my always-on Internet connection, you won't be able to return the call - unless you know my Internet IP address. This may, however, be a very temporary situation.

The Business 2.0 analysis also fails to take account of Bluetooth. It's understandable that it would, because its report points out (correctly): "Also, 3G networks (like Verizon Wireless's 1xRTT Express Network) provide coverage over huge geographical areas, much like present-day cell phones. WiFi systems, on the other hand, have a very short range -- generally less than 300 feet."

It sounds logical enough; and on the face of it, Bluetooth restricts the user even more; to probably no more than 30 feet from an access point, using most current technology.

This (the report suggests) effectively forces WiFi users to remain in one location while connected.

The trouble is, it doesn't. Today, a WiFi device can at least roam around an office building; and any VoIP calls it is carrying will continue as long as the user is on-campus - quite possibly, out into the street for a few feet. But a more important point is the fact that most of the time, the user probably isn't forced to remain on-campus; they choose to remain there.

Already, a huge proportion of "mobile" phone calls are received in offices, and quite a large proportion are made from them, too.

But both WiFi and Bluetooth can be built into hand-held mobile phones - and Bluetooth is already fitted as standard to a dozen or more designs, with more to come.

The killer, of course, is that you still can't receive calls. The mobile phone companies are in control of phone numbers for mobile phones. You might well assume that the one thing they won't do, ever, is give numbers to Internet Providers who will steal their revenue by routing the calls using VoIP - and you will probably be correct.

But there's nothing to stop them selling these numbers to Internet companies like CallServe. And so the question really is: "Would they be prepared to sell them at a big discount?" Well, maybe they would. Here's why:

Already, ordinary GSM phone services are congested, and forecasts suggest that their coverage will become saturated by 2004 or 2005. The huge rise in "texting" -short message service, SMS - accounts for a lot of this congestion. But so is voice traffic.

To increase coverage in GSM technology will be enormously capital-intensive. New masts and the lines to support them are harder and costlier to install, as public resistance increases - and installing 3G networks is even harder and more expensive. But if Vodafone, say, could shift half its GSM traffic onto the Internet, for a flat fee paid by an Internet provider - why, then it could postpone its capital expenditure for a long time.

More to the point, with spare capacity on its network, this operator could easily provide a killer GPRS service.

Today, GPRS uses spare capacity between voice calls to transmit raw data between phone and Internet. But effectively, because urban congestion is so acute, there is no spare capacity. So although, in theory, you could see data transfers running at 120 kilobits per second, reality is that you will be lucky to get 14 kilobits, and even that is going to be available only for short bursts.

Shift the congestion sideways, and suddenly, your customers will be able to use genuine 2.5G data connections, without being charged absurdly prohibitive prices for megabyte. You would become hugely competitive. Yes, the revenue per call over Internet would be tiny - maybe 10% of what it is over GSM, maybe less - but you might gamble that the increased data revenues would make it well worth while. And, of course, your remaining voice customers would get a noticeably better service, with fewer "Net congestion" messages and dropped hand-offs.

Right now, the only thing stopping this from happening is probably the fact that there isn't a big enough VoIP provider to do a deal.

Developments like Sputnik may change all that. Sputnik sets up WiFi access points anywhere, for the cost of a used PC and a £90 WiFi card - and creates a huge, franchised network linked via the Internet. You will, if it takes off, be able to sign up for a monthly fee rather smaller than a cellphone subscription, and have access to the Internet wherever there is a Sputnik satellite - and if you have your own satellite, it would be free.

Sputnik doesn't support Bluetooth, yet. If it ever does, that will be the beginning of a big raincloud over the 3G parade. Because if Vodafone (or any other carrier) were smart enough to do a deal with Net2Phone or CallServe and Sputnik, assigning phone numbers to Internet-based users, then suddenly, any Bluetooth-equipped phone could switch to IP for both incoming and outgoing calls in home or office, and only use the GSM network out in the countryside where its longer range is needed.

In any event, it could all become academic within two years; the one factor that would free the Internet from any need to use phone numbers, would be the adoption of the new Internet protocol, IP version 6.0. That isn't here; and it's been "coming next year" for as long as most Internet watchers can remember.

If we did have IPv6, then phone numbers would become obsolete. Every phone, every PDA, every information appliance you can think of, would be able to have its own unique IP address. Connecting a call to such a device would be possible with no more than the IP address - it would be come, effectively, the phone number system of the future.

Well, there have been many false dawns; but talk to the makers of Internet test gear, like Spirent, and suddenly, 2002 looks like the year IPv6 may happen after all. It won't happen for six months, of that, they're sure. But it's definitely moving, at last; people are making IPv6 equipment that will need testing, and they're ordering the test gear.

"We first saw this some months ago, in October," revealed Spirent's VP of technical strategy, Mark Fishburn, at the NetEvents seminar in Montreux in early March. "Up till then, we had the normal vague statements of good intentions from people like Cisco and Nortel, but no interest in spending money. Now, it's suddenly different."

There is still some way to go, from the theory, to the practice. It may be theoretically possible to use a Bluetooth phone to place a VoIP call through a Bluetooth access point - but right now, nobody provides a system to do this for the ordinary user. And one day, WiFi circuits may be produced with the low power consumption needed to work in mobile phones; but today's WiFi phones are big, and heavy, with powerful batteries to cope with the current drain, and would not sell well in competition with neat, fashionable GSM toys. Most significantly, there is not a global VoIP provider with access to the right number of WiFi or Bluetooth users, to attract the attention of companies like Cellnet or Orange.

And finally, there is still not a real understanding, inside the GSM companies, of what they could be doing with WiFi and Bluetooth, if they chose to do it; nor of the very real threat to their monopoly of phone numbers that the Internet version 6 will pose.

But the awareness is rising. If you're going to watch the 3G parade, take an umbrella. I don't know for sure that WiFi and Bluetooth will rain on the parade, but the clouds look ominous to me.