April 21, 2004


VoIP hype and Skype.. Nice background on the service.

COMMENTARY--The music industry rode to a zenith on wax cylinders, plastic discs, tape and CDs--and then along came MP3. The jury is still out on the ultimate impact of Napster, Kazaa, iTunes et al but it looks like a major turning point.

For telcos, many see the arrival of the Internet and Voice over IP (VoIP) as a parallel situation. While there is a lot of hype, exaggerated claims and oversimplification, the risks are real enough. This is an industry with over 100 years of history, an established infrastructure and a dedicated customer base often too lethargic to move. But it is also a highly competitive and deregulated sector struggling with rapid technology change--and it may be staring death in the face.

In the midst of all of this arrives Skype, presented by the founders of Kazaa. This little known service provides a simple downloadable software package that allows PC owners to make good quality telephone calls using VoIP. Users can call other Skype users for free, while connection to the telephone network and other services attracts a charge.

In addition, Mac users have iChat, which includes voice and video for free. Then there are a range of providers, software products and services offering VoIP connections of variable quality.

So, could VoIP become a serious threat to the phone companies? Well against the one billion or so fixed lined telephones and over one billion mobiles, Skype has so far seen 10 million downloads and total VoIP service estimates see less than 100 million users. However, such is exponential growth that industry estimates forecast 40 per cent of all calls will use VoIP by 2007.

Everywhere I go in the United States I now see people with PDAs, laptops and headsets making VoIP calls. This has been compounded and supported by the rapid spread of Wi-Fi providing a very powerful platform for users on the move.

The mode of operation spans the normal fixed/mobile phone behavior, plus the use of e-mail to establish contact and prompt the use of Skype, iChat, etc.. The more adventurous are also linking screens and working cross platform--with common applications and displays--in a manner forecast a decade ago but still seldom seen on corporate networks.

I think it would be foolish for any telco to dismiss VoIP and especially Skype. It seems to me that DIY telephony is on the march and will soon be on the scale of Kazaa.

But we should also remember that there are limitations to the performance of IP networks and thus to VoIP or indeed any real time service. The reality is the Internet is fundamentally ill-conceived and ill-equipped for the support of real time services of any kind. If you try VoIP you will no doubt be pleased by the quality of the voice connection for at least some of the time. The snag is that for a significant proportion of the time you can find that the quality is extremely poor.

When you make a traditional telephone call a direct and dedicated connection links two telephones for the transmission of those bits throughout the call. This is referred to as circuit switching. On the other hand, when you send an e-mail message via the Internet your bits are transmitted in discrete packets that are often routed very differently, packet by packet. So the arrival times of packets can be widely distributed and vary on a second-to-second basis.

For the Internet to rival the telephone network it has to have an over-capacity to ensure that communication between two fixed points can be nailed down and held reasonably stable for the duration of a real time service such as a telephony or videoconferencing.

Is this possible? Yes. What is more, it is possible at about 10 per cent of the cost of the old telephone network. Bandwidth is the cheapest commodity we now manufacture and its provision in IP networks is trivial compared to the telco environment.

Is there enough bandwidth available on the Internet to immediately replace all the telephone capacity we currently enjoy from the mobile and fixed line operators? No. But we are in transition. We are moving from a world of circuit switching to a world of packet switching and we have to invoke network design constraints to make real time services 100 per cent viable.

The most basic and fundamental is constraint is restricting the number of hops between routers and switches connecting any two points. For a successful global telephone network based on IP technology it would seem that something of the order of five hops would be a near ideal target to minimize the latency on packet delivery.

In all of this there is an existence theorem for the tolerance of customers that says: No telco engineer or manager ever anticipated that so many would pay so much for such a poor service as that provided by a mobile phone connection. The fixed line network is one of strict quality control and the maintenance of extremely high standards but 12 years saw it rejected by customers that value mobility above all else. I think the Internet heralds yet another change in customer values as they seek and value a self-determined and unrestricted DIY world of low cost everything.

Is there is a downer in all this? I think there is and it comes in the shape of the virus, Trojan horse, worm and spam. Over 50 per cent of todays in-use Internet capacity is being consumed by these negative activities. For VoIP services to become universal we will have to see some constraint put on rogue activities. Alternatively, we will have to provide and waste bandwidth on a huge scale.

Will all of this happen? I think so. How come? For the simple reason that it can be done and people will demand it. Like Wi-Fi, it has become visible and useful and, what's more, it is becoming desirable and irresistible.

Draft dictated to digital voice file and transmitted to my PA via Wi-Fi from my home. Typed version collected 12 hours later in the BA lounge T4 Heathrow via my 2.5G mobile phone (I would have used Wi-Fi but not at $20 for 30 minutes). Copy revised on BA297 flying to Chicago and final text dispatched via in-room high-speed LAN at a downtown Marriott.

Peter Cochrane is a co-founder of ConceptLabs CA, where he acts as a mentor, advisor, consultant and business angel to a wide range of companies. For more about Peter, see: http://www.cochrane.org.uk/. He writes a regular column for silicon.com that is archived here.

Posted by Craig at April 21, 2004 04:44 PM