The Internet Telephony Dilemma
Dr. Arun Mehta


They say that when you wish to trap a hedgehog, it is a good idea to use a bucket. You first quickly invert the bucket over the hedgehog's ball of quills. Now, goes the advice, you have a place to sit, while you contemplate what to do next. Effectively, this is what the Indian government has done: covered up the Internet telephony problem through a simple ban. Now, however, with the impending privatization of VSNL, the government cannot escape the problem it has been sitting on much longer. Meanwhile the "problem" has been growing ever sharper spikes.

As has been widely discussed[i], Internet telephony is more efficient by an order of magnitude over circuit switched. It also, typically, is not saddled with the huge accounting costs of the conventional network. New services such as telephony-enabled websites and online collaboration often have Internet telephony at the core. The arguments against the ban aren't going to go away, only get stronger.

Meanwhile, allowing VSNL a monopoly over telephony to and from India has fattened this calf, giving it huge profits and reserves ahead of international bidding for a slice of this company. Its net profit in fiscal 2000-01 soared by 87 per cent to Rs 15.78 billion. Total revenue for the year ended March 31, 2001 grew to Rs 78.76 billion from Rs 73.24 billion, registering a growth of about 8 per cent over the previous year[ii]. However, this special treatment hasn't made the company any healthier, as it contemplates existence in an internationally competitive environment, for the company has only shown an ability to make profits in a monopoly situation. Investors are worried about its ability to compete, not just with new competitors expected in 2002, but also with new technologies that threaten the very paradigm of conventional telecommunications that it largely relies on.

With extensive optic fiber deployment, and pressure to sell all the bandwidth these cables provide, wireless broadband networks seem to hold the key for rapidly bringing bandwidth to large customers. Unfortunately, 3G, the suggested approach of the telecom multinationals, is still a couple of years away. Meanwhile, 802.11b has emerged as an exciting alternative, which extends the ubiquitous Ethernet LAN into the air.

802.11b costs less than a hundred dollars a node, routinely carries traffic in excess of half a megabit per second, traverses kilometers with line of sight, uses off-the-shelf hardware and software. Because it simply transports TCP-IP and the Internet into the air, it doesn't need a central server, thus reducing the role of the large telco to deploying long-distance optic fiber networks. Local distribution is far more easily and cost effectively done by the ISP, cable TV operator or small entrepreneur. VSNL's notoriously poor customer relations in any case make it an unlikely winner in the telecom retail market.

Another important consideration weighing on planners is robustness. TCP-IP was designed with disasters in mind. A real-life test of this concept came during the Gulf War in 1991, when the US military faced huge difficulty in knocking out the Iraqi command network, which was built using commercially available Internet technology[iii]. In the event of a disaster, the surviving nodes automatically reconnect, unlike conventional telecommunications networks, which have consistently let us down in natural disasters such as occurred in Latur, Orissa and Bhuj.

In 802.11b, the individual invests one-time in the networking hardware, which depending on distance and the presence of other 802.11b nodes, costs between $40 and $500, but significantly, doesn't pay the local phone company ongoing rentals. Networks as small as 2 nodes are commonplace, and can easily scale up to hundreds within a building. Even within a room, the economies of scale that Bluetooth isn't finding, 802.11b reached by simply extending the wired Ethernet LAN into the air.

For developing countries, very exciting is the community telecommunications model that 802.11b opens up, as demonstrated by consume.net in London and environs (http://www.consume.net/consume-generalfaq.html) and others. Using this model, a group of people can buy bandwidth in bulk, and share among themselves. A dozen people might, for instance, share a DSL connection, where one at a time, each can do high quality video-conferencing! At the same time, others might use the network to share a video, or exchange songs.

In rural areas, this could be used for telemedicine, allowing the doctor in the city to see patients in far-flung communities. Distance education also works best in two-way, high-bandwidth environments.
 
Many people in rural areas are therefore likely to see TCP-IP networks before conventional circuit-switched telephony reaches them: will they be denied access to telephony, merely to support the profits of a semi-private company? The problem for this cash-starved government is that VSNL is the goose that has been laying golden eggs - its short-term interests as a shareholder in VSNL corrupts the ability of government to make sensible policy in this area.

Once upon a time, the telecom business was cosy. The telecom switch was like a tent in the desert. Monopolistic telcos were like Arabs living alone in these spacious tents. Things didn't change much, basically determined by how quickly lawyers could come to an agreement at the ITU. Network intelligence was only at the switch, which the telco had full control over. What the poor camel, also known as the customer, received - a phone instrument - was a dumb terminal. But now, telephone instruments all contain computers. More and more, they have even started to look like computers. Internet telephony was the nose of the camel, seeking to enter this tent. 802.11b is the head, and even the neck. Those telcos that did not try to keep the camel out, at least stood a chance of keeping the tent of their business intact, even though they now had to share. By keeping Internet telephony out, VSNL might have kept the tent cosy a while longer. The danger, though, is that when the camel does finally come in, it will just tear the tent apart.


[i] See for example "Ringing in the Old", Arun Mehta, Business India, October 5, 1998.
[ii] "VSNL profits soar 87%, Internet base up 72%", Times of India, May 1, 2001
[iii] "The Internet for Dummies," John Levine and Carol Baroudi, IDG Books, 1994

Comments