On 1/02/2012, at 10:49 AM, Alastair Johnson wrote:
On 1/31/2012 1:04 PM, Joel Wiramu Pauling wrote:
The general attitude towards peering is that:
Peering is an issue for RSPs to deal with in the context of UFB so "is up to the market" to decide.
This makes sense. As underwhelmed by the ComCom report as I was, I'm not totally convinced there is a significant peering issue.
I'm thinking some kind of statement around market power abuse with regards to peering, but ultimately it is a commercial/market issue and for the most part it seems entirely manageable in New Zealand at this time (and in the immediate-ish future).
I have an inkling as to why this keeps being raised as an issue. The logic goes like this: Step 1: When someone sets up an ISP or large content network the first connectivity they buy is transit by which they gain access to 99.9% of all routes. For resilience purposes they generally buy another transit contract and hopefully they get the other 0.1% of routes through that. That way they have full connectivity and the basics are taken care of. Step 2: If the cost of this transit is sufficiently high then the next step is to look at what routes carry the most traffic and see if those routes are local and can be connected to by cheaper means. One way of doing that is with a direct contract with the network that has those routes[1] but the most common way by far is peering with other networks for exchange of *local* traffic. So, peering is basically a cost reduction strategy whose effectiveness is determined by the cost of transit. (To be clear I'm not saying peering is a way of getting free transit - peering is only about *local* traffic exchange using whatever definition of local the market accepts.) Now we get on to the NZ experience and there are two important characteristics of our transit market to take into account. - We have market segmentation and differential charging between national and international transit, when places like the US and UK just have one product called IP transit that gets you everywhere. - Our transit costs are very high in comparison to the US which is only a cable hop away. This makes peering potentially a very effective cost reduction strategy in NZ. Recognising this, our ISP/content provider from Step 1 sets out to peer locally to reduce their transit costs and find that in NZ that the market looks like this: a. Two of the biggest ISPs don't peer, not even with an extremely tight definition of local, though one of them does offer [1] above. b. Those same two ISPs are major transit providers in a very small market of suppliers. And hey presto, that's where the problem with peering is. What should be clear from this though is that peering is a second-order problem and the first-order problem that should be tackled is the high cost of transit and the ridiculous market segmentation. In the US transit costs are so low that many transit buyers don't need to bother with peering. cheers Jay
If someone turned up with 40Gbps of traffic today between two operators, I suspect those two operators would be only two happy to interconnect vigorously. Just think of all that over-cap-charging revenue!
_______________________________________________ NZNOG mailing list NZNOG(a)list.waikato.ac.nz http://list.waikato.ac.nz/mailman/listinfo/nznog
-- Jay Daley Chief Executive .nz Registry Services (New Zealand Domain Name Registry Limited) desk: +64 4 931 6977 mobile: +64 21 678840