Hacker News Reader

Comment by amazon_not | Hacker News Reader

Bandwidth isn't really free, even to the so called Tier 1 carriers. Transport, crossconnects and ports cost money regardless of if you have settlement free peering or not. More often than not the other costs are far larger than the actual bandwidth.

Even if the so called Tier 1 carriers are losing their prominence due to donut peering of smaller operators and large content providers building their own networks, getting settlement free peering with the big ones remains a challenge. Settlement free peering itself isn't an exclusive club, but being an operator with significant market power is.

While the majority of settlement free peering agreements are informal, studies show that about 10% are more formal with written contracts and conditions.

View on HN

My comment was in response to price charged by one party to another to move bits, bandwidth as in opex. Of course nothing is free and gear is capex, albeit with heavy discounts and handsome depreciations on capital leases.

Even with public peering fabric which is what I think you are talking about and that I've used LINKS, NYIIX and AMS-IX etc,. you still have to backhaul that traffic yourself, its not making a significant dent in Tier dominance.

What studies have shown that there are contractual agreements for settlement free peering? Citation? Links? I would be interested in seeing that. My source was from experience working in the industry.

I'm not saying the cost is significant in the grand scheme of things, but unless you own the whole physical network, the co-lo facility and the IX then your transport, crossconnects and port costs are OPEX, not CAPEX.

Transport, crossconnects and port costs apply regardless if you use a public IX or do private interconnects.

Here's a link to the study about settlement free peering: https://www.pch.net/resources/Papers/peering-survey/PCH-Peer...

I had the percentage wrong, or then I may recall a more recent study. This one is from 2011. Of course it depends greatly upon who you peer with, as to whether a written contract is needed or not. As such the sampling may skew the results.

We (PCH) committed to performing that study every five years, and we're beginning the process again now, for a 2016 edition. Expect results in six to eight months. Based on what I've seen of the industry in the past five years, I'm guessing there won't be any huge change since the last one.

           -Bill

Reply on news.ycombinator.com