The Bandwidth Revolution
Strains Internet Growth
by R. A. Lemos The internet definitely is not dead in Japan. Yet, despite what the hype would have you believe, neither is it a roaring industry. The quasi-limbo in which the Japanese Internet finds itself has a single root cause: leased-line prices. More than just NTT's monopoly, the high prices stem from a great many factors. In the end, providers' inability to offer adequate bandwidth will slow the Internet's growth, and result in providers seeking more profitable ways to fund their connection.
Today's Internet, while still primitive, goes way beyond its humble ARPANet roots. Thanks to the World Wide Web, the Internet now carries information in varied formats: text, pictures, video, and sounds. The Web has made sending and receiving these resources very simple -- almost fun. And this means that more people are doing it.
Receiving a picture over the Internet, however, is very bandwidth-intensive. The saying claims "a picture is worth a thousand words," but in digital format the file for even a modestly sized picture is 100 times that. In the not-so-distant past, a user might have transferred only 100KB of data in a typical session. Today, a typical Web user will transfer 1MB or even 10MB of data.
This is the cause (and effect) of the Web revolution. The Web not only makes the Internet more appealing to more people, but also encourages each user to use more bandwidth. In the US, new applications and ways of using the Internet will be found to profit from this new freedom. In Japan, finding new ways of making money is imperative, as new sources of income will have to subsidize the necessary bandwidth. Until these electronic enterprises start to pay off, bandwidth will remain scarce and overburdened.
A quick fix?
This puts providers in a catch-22 situation. Providing adequate bandwidth requires generating income; but in order to generate enough income, companies need to oversell bandwidth. Many of the providers have jumped to a fixed-rate service, yet in view of the bandwidth demands of today's users, a cheap connection is invariably an overburdened one. To see why, it is important to understand how a provider makes money.
The provider of T1 service (1.5M-bps) may have about ¥10 million in line fees and salaries to pay for each month (a conservatively low estimate). Assuming that a provider is generating funds based only on selling its bandwidth, at a fixed rate monthly fee of ¥5,000, the provider would need 2,000 customers to break even. If even 10% of these customers wanted to connect at the same time, each would receive just 7.5K-bps of bandwidth. ("Bursty" transmissions would have a higher probability of receiving more bandwidth, but the Web relies much more on average download rates. A single packet may come in at 64K-bps, but the download rate is still slow if you only receive one packet every 10 seconds).
Unfortunately, in Japan, this quandary leads to an inevitable trade off: customer satisfaction (bandwidth) vs. profitability. With all of Japan moving up to 28.8K-bps modems, and ISDN making inroads faster than in the States, users expect more bandwidth. The example above assumes a T1 provider, yet few providers in Japan actually use a T1 line. (One of the largest, IIJ, has two international T1s that it leases to 10 or 20 smaller providers.) Most local providers have (at most) a 256K-bps international connection, which can result in glacial individual connect speeds below 5K-bps, and unhappy customers. Yet, to handle fewer customers (and increase throughput) would decrease the company's revenues.
Providing a solution
One solution is for providers to resist the temptation to move to a fixed rate -- and for customers to stop assuming that the cheapest service is "the best." For providers, basing charges on when the service is accessed (similar to the telephone companies' structure of prime, economy, and discount time slots) would help the system police itself. Or, possibly, offering a fixed rate service during slow periods, with surcharges for accessing the system during prime time.
Another solution, and one that many providers are already taking, is adding value to their bandwidth services. Several companies are generating income by increasing their Web presence to create a virtual space that is valuable for advertising. Others are becoming clearinghouses of information: some for free, others as a paying service. Still others are pushing corporate accounts and special services to subsidize their personal users (business and personal usage peaks at different times, which helps to even out bandwidth demand). The number of providers expanding their bases enforces the impression that mere bandwidth providers cannot long exist in the consumer Internet market.
As the Internet becomes more a system for distributing general information in varied formats, the old system of pricing will cease to work. Only after decades of experience with the telephones did AT&T decide to go to a fixed rate system of charges. With the Internet, there is no way to judge the proper prices for services, because services change so quickly. At the moment, users can get electronic editions of newspapers and magazines, hear radio broadcasts, check out new CDs, and (if they can get the bandwidth) do videoconferencing. Yet, at fixed pricing, providers cannot afford the bandwidth needed to provide many of these services, leaving the Japanese Internet stuck in its cocoon.ç
