President-elect Barack Obama has been clear that investment in infrastructure — including Internet infrastructure — will be one of his administration's signature initiatives. That's excellent news, because the Internet could surely use some focus and investment.

Here's hoping, though, that his team brings the same attention to detail to his Internet initiative that it's brought to the campaign, staffing and transition strategies — because fixing the Internet isn't quite as simple as it looks.

First off, there's the definition of what exactly the problems are. In previous columns I've outlined two significant technical issues: the lack of bandwidth capacity at the edges (access lines), and the ongoing structural fragility due to inherent architectural flaws (particularly around naming and addressing).

Those are the technical concerns. But there are others having to do with policy and economics. For example, is Internet access a "right" or a "service"? What is net neutrality, and does it apply equally to carriers and content providers? What is a user's expectation of privacy, and does it apply equally to carriers and content providers? And finally, what is the definition of "broadband" services?

It's all very neat and simple, or so folks would have you believe. But I'm with noted curmudgeon HL Mencken on this one: "For every complex problem, there is a solution that is simple, neat, and wrong." (There's a similar quote attributed to Einstein).

In previous columns, I covered some of the nuances around Internet access, net neutrality, and unequal perspectives on privacy (it's apparently OK for content providers, but not carriers, to spy on their customers — hmm).

So what would I like to see from the next administration's Internet infrastructure initiative? For starters, I'd like a group convened to discuss some of the technical and architectural challenges of the Internet. And no, that's not the Internet Engineering Task Force, which has done some great work but has pretty much passed its "use-by" date.

Thank the IETF participants for their decades of worthy service, and start fresh. Invite some of the folks who've correctly predicted some of the current issues, have operational experience, and have continued to do innovative research. I'm thinking of people like John Day, Mike O'Dell, Noel Chiappa and Tony Li. Task them with coming up with a next-generation architecture — and having it implemented before 2011, which is when many of the current issues will reach the crisis point.