Breaking Through the Glass Ceiling
by Craig Oda, TWICS Co., Ltd.
At the end of June, system administrators from all over the world got entirely
too much sun at the 1995 International Networking Conference (I-NET) in
Hawaii. The conference covered the technical issues surrounding the Internet
as well as business and legal/ethical issues.
The Asian region sports the highest growth rate for the Internet, so it
wasn't surprising that many of the papers concentrated on Asian Internet
issues. In one paper, Pindar Wong of Hong Kong Supernet proposed an audacious
plan to unify the Internet of Asia and bring more bandwidth and Internet
access to the region. Wong warns that most Internet service providers (ISPs)
in Asia either will find it difficult to reach critical mass, or will run
into a glass ceiling.
Crowding the lines
Reaching critical mass is a worry of all Internet service providers. An ISP requires a minimum number of customers to avoid losing money. Only when an ISP's income, generated from the fees of its customer base, meets or surpasses its monthly expenses can the ISP be said to have reached critical mass. The exact cross-over point into critical mass depends on the country and ISP's niche, but the monetary amount is definitely higher for Asian ISPs than for their US counterparts.The largest expense component for a start-up ISP is the cost of Internet bandwidth. Wong estimates that the cost of international bandwidth in Asia is, on average, 50 times more expensive than that available in the US. Internet bandwidth is usually sold in 64K-bps units, with the typical start-up using an ISDN leased line to a larger telecommunications carrier. Because of the much higher monthly cost of their Internet connections, to reach critical mass Asian ISPs must either charge their customers higher prices or get more customers. Unfortunately, in most Asian markets, competition has forced down prices; in general, demand for high-priced individual accounts has disappeared. Thus, ISPs in Hong Kong and Tokyo are forced to crowd their bandwidth with a larger number of customers in order to reach the break-even point of critical mass.
Breaking the glass
Even after an ISP reaches critical mass, it still faces the obstacle of a "glass ceiling" before it can stabilize and compete effectively with other large ISPs. To grow from a "garage" business into a viable commercial organization, an Asian ISP needs a T1 (1.5M-bps) leased line to the Internet. (Only at the T1-level does economy of scale begin to work; anything less than T1 forces the ISP to pay much higher per-unit prices for bandwidth than its larger competitors).A technical T1 for an ISP in the San Francisco Bay area, for example, although it comes with very little support, goes for $800 to $1200. A T1 in Hong Kong goes for US$80,000, and one in Tokyo goes for US$60,000. The price of a T1 thus becomes a barrier that needs to be passed in order for an Asian ISP to survive -- the glass ceiling that prevents many Asian ISPs from expanding.
In Japan, this problem seems to have been overcome for some ISPs through a mixture of venture capital and joint-venture agreements. Many of the smaller ISPs in Tokyo will encounter this glass ceiling, though, when they try to expand and compete with the larger companies.
Bringing bandwidth to Asia
The solution that Wong proposes to overcome the cost of a T1 connection is intriguing: his strategy is to buy more. Wong claims to have been contacted by providers in seven different Asian countries asking if Hong Kong Supernet could supply them with Internet connectivity. This has prompted him to push for an Asian Internet infrastructure that uses Hong Kong Supernet as the hub. In this scenario, most of the intra-Asia Internet traffic would stay within Asia, without going to the United States. (Currently, although it makes sense to directly connect Hong Kong with Taiwan, in reality all traffic goes from Hong Kong to the US, and from there to Taiwan. Information to be sent from Japan to Korea, likewise, goes via the US.)It is unlikely that Wong's project will be viable in the near future, and the selection of Hong Kong as the Asian hub is questionable. (The question of what happens after 1997 makes Hong Kong a risky proposition, at best.) The concept
Craig Oda came to Japan in the summer of 1990. He met Tim Burress, the current president of TWICS, at the Japanese Language Institute, and was at TWICS in the fall of 1993 when it became the first public-access Internet system in Japan. TWICS also markets the Global SprintLink Internet service using direct IP, frame relay, and international leased lines. Craig can be reached at craig@twics.com; his personal home page is http://
www.twics.com/~craig/home.html.
