More on IP transit plus peering

We’ve had a very fruitful running conversation about traffic growth and the wholesale IP transit market amongst Ike at Telecosm and Dan at BearOnBusiness.  Today Ike has an excellent post about why the IP transit market is what it is today, and how owning the fiber itself is going to matter more as growth continues.  Personally, I think that there will always be some room for the Cogents, the Hurricane Electrics, and the Verio’s in the IP transit market, but I do agree that the landscape is beginning to change on them.

In a post the other day, I described several reasons why the wholesale IP transit business is not a good proxy for internet traffic as a whole, commenting on the effect of CDNs like Akamai peering at the edge, differing geography, and differing levels of competition.  As several comments noted, another major effect is the growth of peering in general – in fact I totally neglected this contribution and I thank them for pointing this out.

Now, tier-1 internet backbones are defined by their settlement free peering with each other, but of course any network can peer with any other.  And when two networks exchange traffic over a peering connection, neither buys IP transit for that traffic.   Of course there are costs associated with peering just as there are with buying wholesale transit, if there weren’t then everyone would peer with everyone else everywhere, right?  It is the relative cost that drives the economics, and it turns out that as the wholesale IP transit market has stabilized and price declines have moderated, the costs to peer have continued to drop.  The result is that the wholesale IP transit market is losing marketshare, i.e. the traffic that goes through wholesale IP transit networks is growing more slowly than the overall amount of traffic as more goes across public and private peering connections.

Private peering is hard, if not impossible, to track quantantatively, and thus people don’t talk about it much and are often somewhat blind to its evolution.  But while it is less prevalent in the USA there are public peering points in Europe that have become huge features in the internet landscape.  The commenter ‘Keys’ the other day left several links to graphs of traffic growth at peering exchanges in Amsterdam and Frankfurt. It is obvious from the growth shown on these graphs that public peering traffic in Europe is growing faster than the IP transit market, and I see no reason why this won’t continue, even after Cogent’s latest pricing gambit.

If you haven't already, please take our Reader Survey! Just 3 questions to help us better understand who is reading Telecom Ramblings so we can serve you better!

Categories: Internet Backbones · Internet Traffic