Todd Underwood made his last post for Renesys the other day, with a perceptive yet dismal view of the wholesale bandwidth business - I recommend a read. Wherever Todd winds up, I hope he finds a way to keep blogging, his voice on the subject of internet traffic has consistently been one of the best around.
He's right though, the wholesale IP transit has been a wholly unrewarding place to be for the entire decade. Pricing remains brutal and differentiation from the competition is nearly impossible. He sees potential salvation in improved quality differentiation, in finding metrics that let the buyers of bandwidth measure accurately the value offered by various networks. I doubt it, the IP transit market seems broken in so many ways that we ought to just never expect to do anything but break even on it. Hunting for differentiation in a sector this broken just isn't the best use of capital. Carriers need to find other services to sell on top of it or alongside it, or not sell it at all.
That's what underlies the CDN pushes of Level 3 and AT&T and now Verizon. It's how small metro providers like Zayo and AFS can thrive selling bandwidth. Depending on the recovery of IP transit as a business model has been the surest way to lose money since the dot com crash, we have to move on.