Downtown Interconnects

This might be my naivate, but a recent Slashdot article brought to mind an idea I’d had back at a technology startup I used to work for.

The whole idea behind some of the basic technologies that the Internet is built on is that there is no center. But everyone who actually uses the Internet is a client of an ISP—they’re “downstream” from the “backbone.”

But the more I see and understand about how things work, the more I wonder why businesses in a metropolitan center don’t interconnect. A single dry pair with a T1 link between two businesses with competently configured routing would effectively give both businesses the benefits of a redundant network link.

If you had more businesses, and arranged the connections between them intelligently, you’d come to the point where only a complete failure of all links out of the business district would sever any of the participating businesses. Granted, we’re assuming mutiple carriers into the participant’s networks, and that their service agreements don’t prevent the plan…

And while technically, it would be quite feasible, commercially perhaps not. Consider how difficult it is to put together a simple peering agreement between two companies whose business it is to work with data networking. How do you seriously propose to arrange for a dozen businesses whose competencies lie outside that arena to arrange the transactions, much less manage the service. (Bringing in a broker company might or might not help.)

At some point this does become practical. When businesses are getting smaller, when each company does a particular thing and does it well, and they need to communicate quickly with the companies across town, not with the head office across the globe. When network links become faster, and more ubiquitus, and easier to come by. And that time might be coming soon.

At least, I’d like to see it.