Telcos And Carriers Forming New Federated CDN Group Called OCX (Operator Carrier Exchange)

Lately, there has been a lot of talk in the industry about carriers and telcos getting together to connect their CDNs with one another, thereby creating a "federating" of CDNs. While this idea has been circulating in the market for many years, with little to show for it, carriers now appear to be more serious about the idea. Over the last few months, I've spoken to a group of carriers who are the founding members of a new group called the Operator Carrier Exchange.

While none of the founding members want me mentioning their company names yet, the OCX is getting close to making their initiatives public. The idea of the OCX is for carriers and telcos to share ideas, come up with some CDN standards and allow one another to connect their CDN networks, in a yet to be determined fashion. As we heard from carriers and telcos directly at last month's Content Delivery Summit, they now want to compete with traditional CDNs and take control of delivering video across their own network.

In principle, the idea of carriers taking over more control of video delivery makes a lot of sense since carriers own the last mile and right now, more than two dozen carrier and telcos around the world are actively building out their own CDNs. BT, France Telecom, Telstra, Telecom Italia, Telefonica, KPN, TeliaSonera, Polska Telekom, Bell and others are all starting to spending some serious CAPEX on their CDN initiatives and others like Verizon are already far along. I've heard some suggest that the carriers have been talking about building out their CDNs for years, so this is nothing new, but the truth it, the industry and Wall Street has been talking about it, not the carriers.

Three years ago, multiple carriers were out in public saying they didn't want to build out their own CDNs as the market was too small, a message they reiterated in 2009 and part of 2010. Instead, carriers decided to partner and resell solutions from the traditional CDNs and spent very little money on any kind of On-net buildout. So while we have been hearing for years that the carriers were going to dominate the CDN market, that message has been coming from Wall Street, not the carriers. Within the last few quarters, carriers are now delivering that message themselves, of getting into the CDN business, but one can't say that carriers have been talking about this for years and done nothing.

Exactly what model the OCX will put forth in terms of how their federated CDN model works is still not known. One model could be where the carrier has a direct relationship with the content owner and then uses the exchange to get access to many other CDNs and a wider footprint. Another model could be where one carrier or a group of carriers act as a broker for the CDN services of every company in the exchange. However the exchange works, carriers will obviously have to have some kind of bilateral agreements with each other and there are many different kinds of interconnection models that could potentially work.

Of course all of this makes sense on paper, but the carriers and telcos have to execute it properly in the market in order for it to work, which is no small task. I would agree with many who would say that the carriers won't move quickly and can't compete with the traditional CDNs in the short-term, but long-term, this is where the market is moving. It's part of the reason why Akamai is developing a licensed CDN offering (LCDN) and why companies like Cisco and others are speaking out on what they are seeing with regards to a federated CDN model.

Some have suggested that CDNs will never work with one another as they see each other as competition, but that's not usually the case. One thing many people forget is that carriers and telcos, for the most part, are regional and only care about their end-user eyeballs. They don't need to be global and build out their own footprint around the globe and this is where the federation model would allow for a larger footprint without more deployment. Also, this would reduce their local transport costs, complement IP transit and peering relationships and allow a regional carrier to sell Internet wide delivery.

As Cisco pointed out during their keynote at the CDN Summit last month, any federated CDN model would have to include a request-routing system, a way to distribute content from one carrier to another and the ability to move accounting data from one carrier to another for billing and reporting purposes. This is no small task, but it can be done. Part of the advantage that carriers have is that they have been working with or reselling CDN solutions mostly from Limelight and EdgeCast Networks for a couple of years, which has given them a good amount of time to get insight into the CDN business and learn about the solutions they will need to provide. That's a big advantage they have since they aren't entering the market blind.

So what's the short and long-term impact of a federated CDN model? In the short-term there is not much of an impact on the traditional CDNs business or market share. For most of them, it's still business as usual, but some like Akamai are thinking about how they might be able to leverage their software to help carriers building out their own CDNs. Long-term, meaning 18 months from now, the CDN market is going to look very different and we will have more carriers competing directly with CDNs. Just look at the deal HBO has with Verizon for streaming and it's clear that more content deals like this will happen in the near-term.

As for the OCX, it's too early to know if they can bring to fruition what their ultimate goal is, which is to try and sign up about 20 carriers to join their federated model. If they do happen to accomplish that, and I hear that they already have half that number of carriers who have agreed to it in principle, those 20 carriers combined would control roughly 85% of all the eyeballs around the world, which would change the current CDN landscape, for video delivery, quite a bit. It's too early to bet if they can do it or not, but with or without the OCX, make no mistake, carriers will play a much bigger role in the delivery of video in the near-term.

  • We also think that federation will be a huge trend in the CDN world. We recently published a white paper sharing some of our thoughts on how this can work:

  • As always, the question is ‘who will operate the master mechanism that routes content to the best server for a request?’ on this ‘federation of telcos’…. I will be very impressed if the telcos get together to do this themselves. I will be moderately impressed if they ‘hire’ a CDN to do it for them, and I will not be impressed or surprised if absolutely nothing happens and CDNs continue to be in business doing the same thing they always have, but maybe pay a bit more to the telcos who have strong last mile user bases. Why build out when a CDN will do it for you?

  • Hi Steve, when two CDNs federate in a proper way they can still do autonomous content and viewer routing. One CDN may not control another CDN.
    There is no ‘master’, it’s more like a hierarchy model where the content publisher rents resources on CDN A and CDN A rents resources on CDN B. (and C, D, E, N).
    If CDN A wants to offload traffic for a specific account in CDN B’s topology region, it can use the storage and delivery resources that CDN B made available for CDN A. CDN A can then redirect viewers to CDN B who will redirect viewers to its delivery nodes.
    This way both CDNs stay in control over their own resources, capacity and the autonomous status of each CDN is respected. CDN A should not be able to claim resources from CDN B or vice versa, without any human intervention. I’ve seen some thoughts on this but it is a bad idea, not just because of capacity / performance guarantees, but also business wise.
    The real trick is to abstract federation from the physical delivery nodes. It should not be locked into specific appliances, or vendors but via open standards based API’s. To abstract federation from physical distribution mechanisms (caching isn’t the only or the best way, should not be mandatory, federation needs to support live relaying, and managed object push as well). To abstract from redirection technologies (not limited to DNS, DNS should not be mandatory: federation needs to support controlled protocol/application layer redirect as well). This is to make sure that multiple CDNs can actually interact.
    In addition you need a set of CDN metadata exchange, e.g. how many resources are available (storage, bandwidth, servers, protocols), what kind of redirection engines are supported, which distribution mechanisms are supported. Access logs need to be exchanged so both CDNs can share (and bill) for traffic usage.

  • Stef,
    Sounds nice on paper, but when has this ever happened?

  • InvestorGater

    In light of the coming federation expected to compete against Akamai and its retinue, it’s good to know that Akamai has this exclusive partnership with the telecom Ericsson:

  • @InvestorGator: That partnership with Ericsson has nothing to do with federated CDN amongst carriers and telcos. The partnership with Ericsson is specific to mobile and as of now, no details have been released by either company on exactly what the service offering will be or when it’s deployed in the market.

  • InvestorGator

    @ Dan
    Okay, thanks for filling in the blank for me.

  • @Stef
    You nailed it…

  • Still waiting to hear of an example of telcos “federating in a proper way”

  • Jonathan Tombes

    Useful update, as always, Dan. But confused on a near-term, long-term question.
    Here is one of your points: “I would agree with many who would say that the carriers won’t move quickly and can’t compete with the traditional CDNs in the short-term, but long-term, this is where the market is moving.”
    And here is your conclusion: “It’s too early to bet if they can do it or not, but with or without the OCX, make no mistake, carriers will play a much bigger role in the delivery of video in the near-term.”
    So does playing a bigger role in video delivery not equate with competing against traditional CDNs?

  • Such a relationship between carriers and telcos would be much more complex technicaly than existing peering relationships. Stef outlines a number of important technical requirements, but methinks it is more complex than he has indicated, given the inherent complexities of http adaptive log processing.

  • Hi Steve, we are doing several pilots on CDN federation. But true CDN federation between multiple CDNs from different vendors as I described is not available yet. It’s work in progress. Everyone talks about it but no-one really has a complete vendor independent solution, it’s either vaporware, or extremely limited.
    The challenge today isn’t actually defining all the requirements as I pointed to above, the challenge is to standardize. Which is quite an issue because I know of at least six (!!!) initiatives around CDN interconnectivity standardization… ETSI, IETF, OCX, EUOCX, etc… as many standards as there are CDN approaches means no standard.

  • Jonathan Tombes

    Clarification: I meant to say that “I’m confused on a near-term, long-term question.” Didn’t mean to indicate that this update was itself confused. There’s likely something that I’m just not getting.

  • The problem isn’t standards- the problem is money. I cannot see any telco sending a time of revenue or allocating cost to any one of their competitors to help enhance the quality of user experience on a competitors network. I’ve been in this business a lot time, and have seen every possible method of partnership, architecture, etc to ‘unite the telcos’ and have never seen a business model that would fly… and still do not.
    Why would cable modem company A ever help benefit cable modem company B in any way when they are competing for subs? Makes no sense at all.

  • Steve, the likely model is not “vertical” federation of competitors within a geographical zone. Rather, it is end-to-end or “horizontal” federation of players in different countries/regions to create a more complete, more global coverage footprint.