Blog Post

The problem with Apple’s peering is that we don’t know if it’s a problem or not

Stay on Top of Enterprise Technology Trends

Get updates impacting your industry from our GigaOm Research Community
Join the Community!

Apple(s aapl) is reportedly building its own content delivery network and is in the process of signing peering agreements with the big ISPs, according to Dan Rayburn, an industry consultant and analyst. Rayburn uses Apple’s apparent willingness to sign peering agreements with ISPs as a way to argue that Netflix’s complaints about having to pay Comcast an interconnection fee are dubious.

At a time when interconnection deals are getting so much exposure, Apple hasn’t used it as an opportunity to argue about the current business models of how networks connect with one another. Much like Microsoft, Google, Facebook, Pandora, Ebay and other content owners that have already built out their own CDNs, Apple appears to see paid interconnect deals as simply part of the costs associated with building out their own CDN network. To date, no other content owner or content syndicator that has built out their own CDN has complained of the current business models or argued about doing mutually beneficial interconnect deals between networks.

But what’s dubious is not Netflix’s complaints over paid peering agreement or Apple’s willingness to enter into these agreements. Instead, the bigger issue is the secretive nature of how content is exchanged on the internet at a time when a number of the significant content and broadband players are consolidating. Also, I’d quibble with Rayburn’s characterization that no one speaks up about these deals. In some situations, such as Google(s goog) or Facebook(s fb), many ISPs happily host content caching servers for those companies without charging them. Cablevision and Sonic.Net both have Netflix Open Connect Boxes, while other ISPs host boxes for Facebook and Amazon for free.

In some cases providers do pay to host their servers inside the ISP, a topic that has been the subjects of government raids in Germany and investigations in France.

One might also point out that ISPs often don’t see Google’s YouTube or even Apple TV’s offerings as the potential disruption to their pay TV revenue that Netflix is. Rayburn points out that Apple TV is a mere 2 percent of total bandwidth during prime time compared to Netflix’s 34 percent, and then tries to point out that when Apple releases a software update it can be huge from a traffic perspective. That’s like trying to justify a crowd containment strategy by comparing the hordes of Black Friday shoppers at Wal-Mart to a normal day.

But, once again, the real issue here is not the differences between Apple and Netflix and whether they should pay for interconnections or not, but that these deals happen in secret when they happen, and in many places they happen in in markets that aren’t competitive. Both edge providers and ISPs consolidating their power — Google’s YouTube potentially buying Twitch, which is a growing source of traffic (about 1.35 percent of total bandwidth according to Sandvine) is an example of this consolidation on the edge provider side. Meanwhile, while AT&T’s(s t) Dish deal or Comcast’s(s cmcsa) proposal to buy Time Warner Cable(s twc) shrinks the overall ISP market — fewer players are at the negotiating table.

This is an issue that the current FCC Chairman has signaled will be an important one for his commission, explaining that while peering agreements aren’t part of the net neutrality debate (they don’t happen at the last mile, but at the point where traffic enters the last mile) they are essential to his “network compact” and the functioning of the internet.

Thus, I’m waiting for the commission to start demanding data from all companies who have a stake in this issue: those that haven’t complained publicly; those that have complained through proxies; and those that apparently are happy to accept the status quo. Until that point, I think it’s a stretch to try to compare Apple’s apparent acceptance of paying for direct interconnections to Netflix’s frustration with the practice. From here it’s like comparing apples to oranges. To find the right comparisons we’re going to need actual data, and then we can figure out if Netflix is just whining or is instead a canary warning us about an important connection point for the internet getting choked off by ISPs.

12 Responses to “The problem with Apple’s peering is that we don’t know if it’s a problem or not”

  1. Michael Elling

    The table Dan uses is a series of apple and oranges. Level3 is not the same as edge access providers (who happen to have their own transnational transit networks). It is a core transit provider like Cogent. Likewise Akamai and Limelight are not content or app providers/aggregators. They are CDNs that either own their own transit or buy from the competitive core transit providers and then cache content (mostly ads delivered) at or near the edge.

    We should be asking why is the transit or CDN market becoming non-competitive? I say this because only in a non-competitive market does an app provider/aggregator feel compelled to vertically integrate and build their own delivery networks.

    As for Dan’s cost argument, none of the content or app providers particularly care about latency; even Netflix. But they do care about transit costs and they are in the best position to decide on layer 1-3 tradeoffs than the balkanized edge access providers, none of whom own more than 40% of the eyeballs in any given market.

    As I’ve said elsewhere, this is not an issue about current supply/demand, rather where should the WAN/MAN demarc be in a world of 4K video, 2-way HD collaboration, seamless mobile BB, and IoT. All of these need cloud economics moved to the edge and the edge access providers are merely trying to prevent that from happening. Been going on for 100 years in one fashion or another. The internet was an accidental result of one of these; namely flat-rate expanded calling areas by the Baby Bells.

  2. The general public relies on journalists to explain things to us that we might see as secretive. It’s fortunate that people like Dan Rayburn are willing to do the heavy lifting that journalists are no longer willing or able to do in the dysfunctional world of click-driven reporting.

    In fact, a perfectly neutral last mile does not go very far toward ensuring a neutral Internet. The differential quality is already achieved before traffic gets to the last mile. In fact, the last mile is the only part of the Internet that’s remotely neutral today.

  3. Thanks for the write-up. Heads up you mentioned “At&t’s Dish deal” instead of AT&T’s DirecTV deal. Points are all there so thanks for the comparison and perspective.

  4. Jack C

    Yep, exactly!

    On top of everything you said, the most compelling concerns over peering arrangements are their impact on new, upstart, and disruptive companies. The fact that the richest company on the planet doesn’t have a problem with the status quo is hardly surprisingly and utterly meaningless in this context.

  5. Interconnection is simple, if network A is dumping inordinate amounts of traffic on network B, network A needs to compensate network B for the burden being imposed on its network. Netflix is audaciously attempting to force all broadband users to shoulder its ordinary course of business operating expenses. That would be unfair and totally irrational.

    • Net Neutrality Sideshow

      Yep. I love how Netflix has gotten the internet mob to create noise for them by conflating these issues. Stacey did a poor job in trying to defend Netflix and then tried to deflect onto the private nature of the agreements. Do you know what the details of Akamai’s deals with Level3 are? Why would these agreements between businesses need to be public? The answer is they don’t. Because they are private business agreements that have nothing to do with the public interest.

    • Jack C

      Your underlying metaphor relies heavily on the notion that bandwidth is an imminently finite resource (i.e., use is within sight of capacity). Sure, traffic in Washington DC is awful. If Netflix was responsible for 70% of the rush hour traffic on a four lane beltway, it would be a problem. However, if the beltway actually had 10,000 lanes, that 70% of 4 lanes is absolutely trivial.

      The added complication here is that horizontally integrated ISPs have overwhelmingly compelling anti-competitive incentives.

      Making broadband internet a Title 2 common carrier (utility) would be a good start. It, combined with other regulations and incentives could spur local broadband competition to ensure that capacity is never challenged.

      • Whatever the capacity is, it didn’t just appear out of thin air. If a private company spent a billion dollars building a toll road, you don’t get to say, look, there are plenty of unused lanes, so let us drive for free.

  6. stuff to think about … yet again, everyone still faces the basic ethical questions they always have and whether or not we are going to address those this time around and resolve the common problem once and for all is up to each one of us in the moment we need to actually do something for the reality we want to create.