My guest blog at GigaOm deals with paid peering and the net neutrality regulations, How Video Is Changing the Internet:
But paid peering may be forbidden by Question 106 of the FCC’s proposed Open Internet rules because it’s essentially two-tiered network access, Norton points out.
Paid peering illustrates how hard it is to write an anti-discrimination rule for the Internet that doesn’t have harmful side effects for all but the largest content networks. Paid peering is a better level of access to an ISP’s customers for a fee, but the fee is less than the price of generic access to the ISP via a transit network. The practice of paid peering also reduces the load on the Internet core, so what’s not to like? Paid peering agreements should be offered for sale on a non-discriminatory basis, but they certainly shouldn’t be banned.
There’s another good treatment of the subject at Digital Society, inspired by the same conversation with peering maven Bill Norton.
UPDATE: There’s an incredible whine-a-thon in the comments to this article by Google’s Director of Network Operations Vijay Gill and some of his friends from a network operators’ IRC channel. Gill says I’ve got all the facts wrong because paid peering existed in a very limited way ten years ago under a different name. I don’t dispute that, but simply note its potential problems with net neutrality regulations in some guises. The issue is whether the Internet of the Future will be a slave to the Internet of the Past’s supposed insistence on a single service level for all peering agreements, not that there ever has been such a regulation.
UPDATE 2: One thing I definitely was unclear about is whether Arbor’s estimates of traffic growth, 47%, are in line with the MINTS estimates. I conclude that overall growth is much higher than the MINTS figure because Arbor measures only inter-domain traffic at Internet Exchanges. There’s obviously been a great deal of growth in the Akamai and Limelight CDNs, neither of which is measured by MINTS or Arbor, and growth in private peering (paid and unpaid) as well. MINTS measures more than public IX traffic, yet their figures are in line with Arbor’s data from public sources only; this difference in method and similarity of measurements suggests that MINTS may be understating the total of the inter-domain case, depending on how the load falls out between public and private sources. Private connections are increasing, according to IX operators and heavy users.
Kittens are about as likely to be forbidden by section 106 as paid peering is. One peron’s whine-a-thon is a pointer at reality for the thinking part of the internet community.
The scope of Questions 106 (and 107) with respect to the proposed Fifth Principle, on non-discrimination, is currently under discussion, so it’s premature to make black-or-white assertions about its final outcome. We’ll give your remark due consideration, of course.