» Articles / News

10/017/2005 Internet: Future Shock, Future Tension
By Thomas Nolle, Network World

Few topics in networking arouse such conflicting emotions as the Internet. Many believe the Internet and its open, fully exploitable model are the future of communications. Others see the Internet's failure to create a self-sustaining profit model as a fatal flaw. Both camps are sort of right, and nothing shows this more clearly than the counterpoint between researchers and carriers on what the Internet should become.

Over the past five years, common carriers have come to dominate IP investment. As public corporations with shareholders to accommodate, this group has demanded that IP pay off on capital spending, as earlier technologies did. In both private network bids such as AT&T's Concept of One and BT's 21CN, and group activities such as the IPsphere Forum, carriers have worked to fit IP into what often looks like the traditional framework of the common carrier.

It's easy to dismiss this viewpoint as a kind of "Bellheads vs. Netheads" tension and therefore dismiss the whole common carrier IP thing. After all, aren't cities such as San Francisco and Philadelphia promoting the idea of free Internet? Isn't free always what the consumer wants? Sure, but it's clearly not always attainable. We don't expect BMWs or yachts to be given away through some obscure ad-financed process. Is the Internet community simply ignoring financial reality?

Maybe, but they're not ignoring the issues of what the Internet has to become. A Massachusetts Institute of Technology research scientist, speaking at a very Internet-centric conference last month, related a story of a colleague's rather insightful criticism of the Internet: "You forgot the protocols that route the money!" He also listed three key things that the next-generation Internet has to address: security, robustness of management and economic (meaning profit) issues. That sounds a lot like the goals of common carriers, doesn't it?

It gets better. There is a research project dedicated to defining what the Internet might become, or should have been all along. The Global Environment for Network Investigation (GENI) defines a series of "partitions" over IP on which researchers can toy with new concepts of IP architecture that might address the three next-generation Internet problems. In effect, GENI virtualizes IP networks, separating service architecture from underlying transport so that many different service notions can be trialed in parallel. Users can select what virtual partition they want to play in.

This sounds a lot like Concept of One, or 21CN, or IPsphere. In fact, with minimal wordsmithing, you could take the presentations of GENI advocates and morph them into something that an RBOC CTO might feel comfortable giving in public. So what's the difference, or isn't there any?

One clear difference is time frame. The research community is talking about an evolution that might take a decade to accomplish. The common carriers are talking about making investments now and aren't willing to wait 10 years to see if a successful model of a future IP service that meets their security, operations/management and economic needs can be developed. It may well be this issue - the sense of urgency at the provider level - that has created a schism between two groups that appear to have common goals.

Carriers' capital expenditure has increased in 2005 and is expected to do the same in 2006. They're investing, and investing in IP. Do we want them to continue to do so? Then we have to solve their problems, and what is really interesting about this research-vs.-carrier tension is that the groups apparently don't talk much. GENI doesn't seem to know about IPsphere, and vice versa.

The Internet is going to be remade in the next five years. The death of many ISPs, slipping profits of many others, increased dominance of common carriers, problems with security and QoS - all of these things are going to be solved. The question is who will solve them, and how that solution will affect what we now call the Internet. Somehow, we have to get carriers and researchers talking constructively if we want much of the current Internet to shine through this future network concept. In the end, money, not innovation, will triumph unless money and innovation can somehow be joined better than they are on today's Internet.