I’m optimistic about bandwidth. It’s too easy to conclude that the Internet is overburdened and in trouble. Though we may see a few hiccups, and maybe even brownouts along the way, there’s a lot going on to suggest that bandwidth and infrastructure will grow more than fast enough to meet consumer demand.
First, telecommunications companies including AT&T, US West GTE, Bell Atlantic, BellSouth and SBC are working on Subscriber Digital Lines, a technology that at least 90% of American homes are capable of using. They all have pilots underway. It promises millions (maybe up to 10 or more) of bits per second to the home or small business. As many as nine variations on xDSL have been evaluated in various stages. It looks like to me like ADSL and VDSL are most likely to achieve large numbers of users. It is really shaping up as a strong race between ADSL and Cable Modems. A standard will develop just as it did for 56K (X.90). 90% of the homes in Holland and 60% in American have cable close by. With that large a market it is a certainty that companies will develop practical cable modems which promise at least 384k bits per second. Time Warner cable has a trial underway that claims in excess of one million bits per second. Standards are evolving in this area too.
Yes, there are some problems with some of these new technologies but my over arching reason for optimism, in addition to the technology, is competitive forces. There are five main players (at least); copper, fiber, satellite, wireless, and cable. This is to the home. None of these has reached it’s potential and each is threatened in some ways by the others. Result will be breakthroughs and leapfrog moves. This is already happening between cable and ADSL. Just like we have seen in the PC industry we are likely to see surprises. I predict we will see quite a bit of this over the next 12-18 months. Maybe it will be wireless that will surprise us. Bits can move through the air as fast as they can move through fiber or wire. CS Wireless and other companies are already deploying two-way MMDS capabilities (multiple megabit capability). The leapfrogging is underway already. Some communities (like Phoenix) have had cable modems introduced and shortly thereafter the telco introduced an xDSL offering. Who would have thought we would have 57.6K bps with analog modems in 1997. We did. Remember when 4800 bps was thought to be the limit for PC modem connections?
The other issue is the backbone; the "superhighway" that links the various hubs of the Internet together. This is where Internet 2 comes in. Over 100 U.S. universities have now said they will build out their infrastructure to include gigapops. Giga is billions of bits per second. Pops mean points of presence.
A new non-profit company called UCAID (university corporation for advanced Internet development) has been formed and it is busily laying out the architecture and advanced applications for the next generation of the Internet. There is much work to be done but the benefits are so compelling I think this is going to happen and quickly (next 18 months). In April 1998 an advanced project called Abilene was announced at the White House which willbe the precursor to this next generation Internet.
The other development of some significance is in the deployment and potential of optical fiber. A recent article in IEEE Spectrum by Alan Eli Willner called "Mining the optical bandwidth for a terabit per second" predicts that 100 billion bits per second systems will come on the market within a couple of years and that terabit-per-second systems should go commercial around 2005. Not really that far off. This incredible increase in bandwidth is being made possible by wave division multiplexing and erbium doped all optical amplifiers. The May 1998 issue of Wired has a story about Qwest, a new fiber optics based Internet technology based telecommunications company. Qwest (and at least four other companies) are busily laying fiber in the ground to provide incredible increases in bandwidth for the backbone. They are laying two conduits in the ground along railroad tracks; thousands of miles worth. One conduit is orange and it will carry 48 fibers for Qwest and 48 for other telecommunications companies. Each fiber can carry 10 billion bits per second of capacity per "window". A "window" is one of 8 spectra that is enabled by dense wave division multiplexing. The result of all this is that each of the fibers will have the capacity for 80 billion bits per second. Times 48! That is more capacity than AT&T, MCI, Sprint, and WorldCom put together!
As for the collapse of the backbone that some people have predicted, Scott Bradner at Harvard points out that earlier in 1996 , a major failure of the power grid left most of the western U.S. trying to communicate by candlelight. He also reminds us that on an average day, more than 30,000 people in the U.S. are without telephone service for an average of five hours each — visit the ATIS Network Reliability Steering Committee page.
I haven’t heard anyone talking about an impending failure of the power or telephone systems.
Another point about bandwidth is that for Intranet applications (inside of a company or university) an organization can have dedicated network capacity at whatever bandwidth they feel is justified. It is becoming more and more common for businesses to install OC12 Sonet capacity (622 million bits per second) for their Intranet backbones. Gigabit Ethernet is beginning to emerge as a technology for incredible intranet bandwidth. Basically, companies can have almost whatever bandwidth they choose (and can afford) to have.
One more thing that makes me so optimistic about bandwidth is that since we all know that having more of it is good, we will be willing to pay for it. Today many people pay 2-5 times more for cable than they do for Internet access. People will demand much higher Internet bandwidth as soon as they know it is available. The marketplace somehow finds a way to get things done at the right performance and price as long as there is competition. Witness the PC market. Since there is intense competition and a free market operating in more and more parts of the world the "invisible hand" will take care of things.
There was also a very bullish story on this in Forbes ASAP magazine (April 7, 1997) by George Gilder called "Fiber Keeps its Promise". George says "Get ready. Bandwidth will triple each year for the next 25, creating trillions in new wealth." More recently the The Financial Times of London had a story on July 9, 1997 called Internet2: Traffic Moves into the Fast Lane.
Epilogue: And while we all complain that 50-100KB is not enough, the Sojourner rover is sitting on Mars communicating with the Pathfinder lander at 2400 baud! See Todd Wallack’s story in Network World (July 14)
