Avoiding the world-wide wait

Despite all the benefits of the Web, the

problem that one often faces is congestion, resulting in slowness of the Internet. While

you get instantaneous access even with a 28.8 kbps modem in the US, I find that effective

bandwidth in India is of the order of a few hundred bytes/sec even with a (shared) leased

line through the VSNL gateway to the outside world. Needless to say, these speeds are not

conducive to the pleasurable activity of surfing the Net.

One of the easiest ways of improving

performance would be to increase the number of caching or mirroring servers in one's

network. Assuming that many users will tend to access roughly the same set of web pages, a

cache, which maintains local copies of heavily-used pages, will increase access speed.

Similarly local mirror servers, which update themselves at off-peak hours, can also be of

great help. But at the end of the day, the network will have to be enhanced.

The main problem with the system in India

is low-bandwidth trunk lines. However, as web usage increases, it is likely that even

efficient systems will begin to run into trouble.

So the simplest solution would be to build

more lines and increase their bandwidth. A number of innovative solutions are also being

proposed. First, is the use of existing, under-utilized networks. Second, is the creation

of new and differentiated networks that offer added value. Finally, there are demand-side

solutions, like differential pricing.

There are also, surprisingly, many existing

fibre-optic cables that have been put in place by various utility companies. Electricity,

water, and gas companies naturally extend their reach to every home, much like the

telephone company. It turns out that many of them have laid fibre-optic cable primarily to

ensure the integrity of their pipes. Clearly, these pipes could easily be used to bring in

the Internet-and so utility companies are looking at offering web services.

In India we, unfortunately, do not have the

same type of utility fibre. But surely, large number of middle-class homes are already

wired with television cable; clearly, cable modems with massive 100+ megabyte bandwidth

could be pressed into service. Similarly, Indian Railways can also lay long-distance fibre

relatively cheap.

In addition to these wired networks,

wireless networks are also being created in the US. For example, in the Silicon Valley, a

company offers affordable wireless Internet access. For $ 39.95 a month and a $ 300

wireless modem, you can get Web connection for 24 hours a day if you wish. Furthermore,

you can access the Web from your car, while you are in a restaurant, and also from a

client's office.

Finally, in the long run, I believe there

will be differential pricing for access to the Web. Those who wish to get faster, easier

access would have to pay more than those who are content with the congestion.

The problems of traffic congestion on the

info highway are effectively the same as those experienced on real highways. In road

traffic too, the simplest and most popular solution to heavy traffic is to build more

roads, so that the traffic can be distributed evenly. However, it can be seen that this is

a losing battle.

Many cities have created car-free zones.

However, this is a bit of a blunt instrument, because it levies a fixed fee regardless of

traffic conditions. Innovations, including smart cards in cars and laser readers suspended

above roads, enable more granular pricing.

Internet planners should evaluate some sort

of demand-management pricing scheme along these lines.

S Rajeev



is Director (Java and Strategic Alliances),


Sun Microsystems, India.