Examining an operator / carrier’s role in Video Content Delivery

I’ve reviewed the impact video is having on networks in articles such as :

One of the culprits in ensuring video is efficiently served into operators’ networks are Akamai, Limelight, CDNetworks, Panther Express, and a host of other CDN (Content Delivery Network) vendors.  We’ve recently seen a number of initiatives coming from Verizon (Partner Port), AT&T, BT (Content Connect) and Level 3, as they build on their existing enterprise CDN experiences to create video CDNs.

The rationale for a carrier-based CDN is they can place the distributed caches deep in the network, very close to the customer, as far as the MSAN (Multi-Service Access Network), home gateway, or even in the STB (Set Top Box).  Their drivers include:

  • Reducing latency of content that their customers request, and operators should know what content their customers like to view;
  • Reduce the demand peering is placing on their network as the content would now be in-network reducing the number of router hops, better controlling the required investment in core router infrastructure as traffic grows exponentially; and
  • Receive revenue for traffic they already transport.  The logic goes: content providers are already paying for delivery and potentially carriers could provide better CDN value as they own core transport and can place caches close to the customer.  Which to be fair, if an operator can regionally or globally compete with Akamai on price and performance, then such competition is good.  Unfortunately this point has been lost in the emotional net neutrality debate.

Traditional CDNs couldn’t get that close without partnering with a carrier and deploying internally (to the carrier’s network), which Akamai has done giving it an impressive global CDN.  However, many of the smaller CDNs connect to an ISP such as Level 3 who then peers with the consumer network operators like VZ, AT&T, Comcast.  Hence there is a valid discussion to be had on internet peering commercial relationships which does not involve the content providers nor the customer.

In the UK tomorrow the government will release another version of its Digital Britain report, interim version is available here.  So BT and the BBC have been posturing before its release.  Unfortunately for BT, the BBC is the media in the UK, so its lost the argument before it even started with the phrase that will come to haunt all operators on this issue “can’t expect to continue to get a free ride.”  Specifically BT has been throttling some of its customers in the evening so BBC iPlayer does not work, and naturally those customers and the BBC are a little annoyed, and BT responded with the above phrase to the BBC’s questioning.

But let’s exame the video CDN market to understand the opportunity for operators.  The video CDN market in 2008 was $400M, of which non-pure-play CDNs like NTT accounted for only $40M.  The market is predicted to grow to between $800M – 1.4B by 2012.  So compared to the $2T telecommunications market its in the noise.  Prices per GB fell by on average 40% in 2008 principally due to Amazon’s CloudFront services commoditizing CDN pricing.   CDN vendors are now offering highly competitive deals to customers, such as for high volumes – 400TB and above – are being priced as low as $0.04 per GB.  Live events may command a small premium (10-20%) over video-on-demand streaming content for efficient bandwidth management with pre-determined peak traffic. Value added services such as encoding, web acceleration and traffic reporting command a similar 10-15 percent premium and are increasingly the focus of the CDN providers.

So examining where this market is going:

  • Pureplay CDN providers will likely consolidate and will likely acquire content management companies to offer additional content VAS (Value Add Service).
  • Operators will partner, acquire and building their own CDNs, e.g. Deutsche Telekom partnered with Edgecast; Level 3 acquired Savvis and the assets of Servcast; and Verizon/BT are building their own CDNs.
  • Net Neutrality will limit an operators ability to monetize CDNs in-country, the business case should be built on operational savings.  Internationally the limitations do not apply.  In-country if they’re more competitive than Akamai, a content provider may choose them, those they’d best act as a consortium as a content provider need only strike one deal with Akamai rather than  750+ deals if its operator by operator.  BT’s statement “can’t expect to continue to get a free ride” will limit all operators attempts to monetize CDNs in-country.
  • Mobile networks are suffering even more than fixed networks, so will likely deploy CDNs for purely operational reasons, ignoring monetizing CDNs.  This will further weaken the position of fixed operators.
  • Given Amazon’s commoditization of the market, value added services are necessary for differentiation such as analytics, trans-coding, and site acceleration.
  • P2P (peer to peer) will play an increasingly important role compared to caching, but the technology and its network impact are still at an early stage, see IETF’s ALTO project for more info.  Again operators have an advantage is owning the network to provide the more efficient solution.

On the issue of broadband pricing being broken, generally operators are making a profit on most of their customers, only a small minority, between 1-3% are using so much capacity that it makes them unprofitable.  If an operator offers a package with a byte cap limit at a discount for most customers, explains why their pricing is broken, and that most customers are unlikely to reach that limit given their current usage habits, this could be a way of realigning pricing to reflect their network economics.  The key is not to take something away without giving something back to the average profitable customer.