What is internet peering then and what’s so desirable about it?
In simple terms, peering describes a scenario where networks close to one another agree to swap data with each other, instead of charging for transit and swapping cheques.
Usually, the swap involves similar data volumes and type of traffic, but it doesn’t have to be exactly symmetric. In some cases, a fee is payable for peering, if for instance a small-volume provider wants to peer with a large-volume backbone.
Either way, peering has several benefits to internet users and network operators. It lowers costs (fewer circuits are needed to connect networks) and provides better performance (peering exchanges use fast switching gear and network paths are substantially abbreviated).
This entirely voluntary peering keeps internet users at every level happy, including techies who have to field fewer support calls thanks to increased route diversity and better performance.
This is all good, of course — unless you are a telco. Telecom New Zealand has never been enthusiastic about peering. TelstraClear told its customers a few months ago that it would cease peering at the New Zealand exchanges in November this year.
Indications are however that Telecom and TelstraClear will continue to peer with each other, as they consider themselves to be the only large Tier 1 providers in New Zealand.
TelstraClear’s move especially has upset plenty of its customers who are either taking their business to the competition (no prizes for guessing which one), or threatening to do so.
Why then would TelstraClear, being the underdog to Telecom in New Zealand, risk such a bad PR move? Rumours have it that the new peering policy is a dictate from Australia, and certainly, TelstraClear techies speaking off the record are horrified at the idea. They have fought tooth and nail to prevent the depeering, which will lead to substantially lowered performance for a large swathe of New Zealand internet users.
The telco rationale is easy enough to see: with peering, the telco cannot sell as many circuits to connect all the different networks. Furthermore, both telcos “own the eyeballs” because they have the vast majority of New Zealand internet customers.
Ergo, telco customers are likely to cause lots of traffic being sent to telco networks simply because they request it from content providers. This, the telco logic says, is wrong, as the content provider should pay to access its customers and for the data it sends, as there has to be capacity to receive it. However, that argument ignores the fact that telco customers are already paying for the data they send and receive, but hey, if you can clip the ticket both ways, why not try?
Peering capers by backbones and telcos are nothing new of course. There are plenty of precedents in the US, Australia and elsewhere – just type in “Peering Wars” into Google and read about the silliness.
Nevertheless, because there is no effective competition in New Zealand for national data transit, the depeering is a tangible threat. There is no way ISPs cannot have been aware that this may happen, so you have to ask yourself why nobody has reacted before?