Netflix Signs Peering Deal With Comcast

February 23rd, 2014 by · 3 Comments

According to various reports, Netflix has signed an agreement with Comcast for paid peering. Of course, this is hardly the only paid peering deal out there and on many levels this is barely news. But it does shine a bright light onto an ongoing shift in how the internet connects to itself.

Netflix has always paid for bandwidth, whether through a CDN, a transit backbone, or whatever. They have been building their own infrastructure, and that inevitably means who they pay for it is shifting. Whereas in the past they might have depended on a Cogent or Level 3 for transit to reach everyone, the more extensive their infrastructure gets the more it can make economic sense to peer directly with some providers. Verizon has offered paid peering to content providers and CDNs for many years. Plenty of transit backbones offer paid peering to smaller backbones, it’s a common enough thing.


But it is *how* Comcast got them to do it that is worrying some folks. Streaming speeds for Netflix videos had been slowing down, not because of interference but because some last mile providers have been resisting upgrading peering infrastructure with the backbones Netflix had been using to reach them. The message is clear: if your traffic gets big enough then you either buy direct paid peering or risk having your transit supplier’s peering points slowly turn into bottlenecks.

Even between major providers, upgrading peering points is a largely voluntary process. The leverage available to get the other side to participate if it doesn’t want to is entirely indirect and can become an exercise in shared pain tolerance. Seeing one network use that ambiguity as one lever to pressure another is, frankly, nearly as old as the internet itself. The somewhat newer thing here is seeing it happen between the last mile and content, with the transit backbones caught in the middle.

But even that is not entirely new, as not so long ago Google solved some problems over in France via a deal with France Telecom that wasn’t so different. But Netflix is Netflix, and the lack of clarity on net neutrality is making everyone a bit more sensitive to such things even when nothing here involves actually interfering with the traffic itself.

Frankly, the balance of power has been shifting away from the transit backbones for a long time. I’ve argued for years that the network neutrality battle would move upstream into peering and transit. Yet if all that happens here is that last mile providers use the leverage of their position to get sufficiently large content providers to pay for paid peering instead of transit, we’ll get off easy.  Transit backbones won’t like it much, but there may not be much they can do about it and the peering/transit world has always been a rough neighborhood.

Now we’ll see if Netflix quickly follows up with the likes of Verizon, AT&T, and TWC with similar deals.  Certainly Comcast isn’t the only one they’ve had problems with streaming speeds lately.  And how big a content provider do you need to be before all this comes into play?

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: Cable · Content Distribution · Internet Backbones · Internet Traffic

Join the Discussion!

3 Comments So Far

  • Anonymous says:

    Rob, I think there’s huge difference between having said something relevant in the past and saying you have “argued it for years,” as you suggest in your blog entry above.

    Your Sept 9, 2013 blog entry entitled “Net Neutrality Gets Its Day In Court, But Does It Matter Anymore?” pretty much contradicts the 2009 entry you reference above to support your claim of a long-standing view on the net-neutrality debate.

    How do you reconcile the two b/c you seem to hold two opposing positions on the subject?

    • Rob Powell says:

      I don’t claim to be consistent over multi-year spans, sometimes I explore different logical paths and sometimes I actually change my mind.

      In this case, I don’t think the positions are as opposite as you think. I maintain that the FCC’s rules were outdated and irrelevant before they were overruled, the battle was already being skirmished at the peering/transit border several years ago with Level 3/Comcast.

      In that 2009 post I was trying to expand the scope of the network neutrality issue upstream. I didn’t succeed at the time, but that peering and transit would eventually be affected is something I have mentioned more than once over the years.

  • Brent Neader says:

    This has blown up a bit, and really isn’t as big of a deal as its getting made out to be. At least here you seem to have the real facts at hand, where many sites do not.

    Direct peering agreements have been happening more and more lately, it just makes sense. You can reduce latency, increase throughput, and normally reduce costs. I’ve seen a big increase in peering between regional ISP’s, for example charter and time warner peer directly when they used to each go through a tier 2 provider.

    As you said Netflix has always paid for bandwidth/CDN, the bulk of it being cogent right now. Cogent has their own horrible history about peering issues (back when level3 actually stopped peering with cogent) as I know firsthand peering between cogent at AT&T has been congested in Chicago for around 8 months. So given congest history with uneven peering stats being already known, Netflix didn’t really pick the best partner did they? This type of one way traffic that Netflix is has made an existing problem much much worse.

    So since there isn’t enough two way traffic (obviously) Netflix will pay Comcast for paid peering, and as a direct result would have to pay Cogent less since the traffic they carry will now be reduced. It’s not like they are paying twice, which some articles out there would lead you to believe. If you read some of the articles out there, they do say that Comcast actually only recently lowered their asking price for this agreement to a point that Netflix could agree with, so in some aspects it’s actually Comcast that caved.

    Given all of that, the end user ISP’s with all the eyeballs will continue to have the bulk of the leverage, so ill agree that that much power (future size of Comcast+TWC is scary big) could be used to cause major issues. In this exact case though as you said, this isn’t new, it’s just on a much bigger scale then we are used to seeing, and has a lot of attention due to net neutrality concerns.

Leave a Comment

You may Log In to post a comment, or fill in the form to post anonymously.

  • Ramblings’ Jobs

    Post a Job - Just $99/30days
  • Event Calendar