Ask DrPeering
DrPeering -
Why bother peering when I can buy transit at $2/Mbps?
Steve Rzenberg
----------------------
Steve -
The decline of transit prices has been going on for a long time and the story is always the same; when considering only the financial benefits (cost savings), peering makes sense when you can peer away a large enough volume of traffic for free to offset the cost of peering. Let’s build a simple example.
Let’s say you are an emerging video player, with a lot of traffic to offload to transit or peering interconnects. Transit. We will go with your assumption that transit can be purchased for $2/Mbps ( today it is rare to find a transit price this low, but let’s go ahead and use this price point).
Peering. Let’s assume that you are already at a populated colocation center with enough traffic to utilize a 10G peering port to capacity, which we will assume conservatively is about 7500Mbps. (Above this capacity you are more likely to see packet loss occasionally at peak traffic, and we want to avoid that.)
Let’s assume that the price for a 10G peering port is $2500/month, a pretty standard price in the U.S.
At capacity, you are paying
$2500 / 7500Mbps=$0.33/Mbps
for that peering port.
Peering has always been a game of traffic volume, although there is also a strong component of skill, leverage, etc. that come into play.
Why Peer when Transit is $2/Mbps?
Tuesday, September 22, 2009
With this simplistic calculation, we can see peering may make sense, especially when the next best alternative is paying transit fees to an ISP in the same building. You could (and probably should) even factor in additional costs, such as local loop fees, peering port hardware, etc. and consider the non-financial strategic interests in controlling your traffic.
Since companies that are pushing this much traffic are becoming increasingly common, we will start sharing our research into the Video Internet Ecosystem in the next few articles.
Dr Peering
