» Dan Rayburn is mistaken in dismissing AT&T as a Content Delivery Network. Mumblemouth:

Dan Rayburn is mistaken in dismissing AT&T as a Content Delivery Network.

Published in December 13th, 2007

Dan Rayburn has some interesting thoughts on Akamai being downgraded because of competitive concerns over AT&T. Not to put too fine a point on it but if Dan hasn’t found AT&T in the CDN business he hasn’t been looking in the right places. When I was responsible for which CDNs to use on the Nine Systems overlay solution there were several months where we utilized over 80TB of delivery from AT&T’s ICDS (Internet Content Delivery Service) product. The majority of this was in streaming services. During 2004 and 2005 we used over 20TB of CDN service from AT&T pretty much every month and it was generally much more than that. We also had to compete against them in large accounts. They certainly had shortcomings but it would be unwise to think that they can’t deliver content. They also have a large sales and marketing infrastructure and are able to leverage their existing relationships with many corporations who also need CDN services in addition to their other services. If AT&T decides to actually make such a move and puts the resources behind it I would worry more about them than I would about Level 3. Though I don’t really think that either is that big a threat to Akamai and I agree that the downgrade was unwarranted. Nor do I think that it would be wise for AT&T to pick up LimeLight. Frankly I don’t think that Limelight has enough compelling, proprietary technology to make it worth it and AT&T certainly doesn’t need the network capacity. They have existing peering relationship with basically everyone. Thus, it’s not as though they’re going to be blocked in delivering content to end users. What they should do is pick up a smaller company that has the expertise that they’re missing. Specifically around creating value added applications to enable their customers to be more successful. Things like Digital Right Management, Commerce, Content Management, Reporting and Syndication. I can think of a couple. The delivery itself is a commodity business these days and margins are being driven ever downward. The value added apps are likely to be the best way to maintain reasonable margins. Akamai had a great opportunity to do this with the acquisition of Nine Systems but frankly I think that they squandered it because it didn’t fit the model that they were used to.
As a disclosure I do own shares of Akamai, which probably aren’t helped by the above.

6 users responded in " Dan Rayburn is mistaken in dismissing AT&T as a Content Delivery Network. "

abe said, in December 14th, 2007 at 1:47 am

Pete,

I will ask the same question I asked in Dan’s blog: How can AT&T manage to place servers in non-AT&T networks as Akamai does? I believe this is an important/necessary feature of any effective CDN architecture.

AT&T can definitely put content physically close to end-users within their own network, but I suspect this will not be readily allowed in competing networks.

pmountanos said, in December 14th, 2007 at 7:34 am

Abe,
Your comment in Dan’s blog was the reason for my reference to the fact that AT&T has extensive peering arraingments. AT&T doesn’t need to put servers inside other networks. Traffic from their servers will flow to endpoints that are on other networks via these peering arrangements. These peering points are distributed around the country with multiple connections to many different networks. In AT&T’s case their peering capacity totals 100’s of Gbps. This is a core tenant to how traffic flows on the internet. In the case of streaming, RTT and packet loss are important factors in performance to end users. AT&T manages their native network and peering to reduce those metrics. As I said in my post AT&T has been doing this successfully for years. All CDN’s rely on the peering of their network providers to some extent. Level 3 is another example of a company that is essentially a core backbone provider utilizing their peering to provide CDN services. Much of content delivered by Akamai will also flow over peering points.

Nick said, in April 16th, 2008 at 6:52 am

Pete,

For the content delivery outside the AT&T’s network, do you mean place streaming servers at the AT&T side of the peer connection with other telco’s network? In that case,the quality can not be better than AKAMAI,cus they put servers in telco’s network even they are on the top.

pmountanos said, in April 16th, 2008 at 8:24 pm

Nick,
I never said that AT&T would have better delivery than Akamai. Nor do they necessarily have to be better in order to be successful. In the case of streaming or large file download they just have to deliver faster than a client can effectively use. The market for CDN services is big enough for multiple providers. Clearly any market with multiple vendors has a “best” but there is no reason that other providers can’t be successful. Take a look at overnight delivery, credit card processing, banking, servers, tattoo shops… the list goes on. Obviously things like app acceleration have different requirements.

Nick said, in April 17th, 2008 at 3:14 am

Pete,

Besides AT&T i know that some top Ts are also planning to build their own CDN.After these top telcos’ enter into this market,what do you think of the CDN market future will be?

another question is, does IPTV system needs those kind of CDN architecture?

thanks

Dan Rayburn said, in January 28th, 2009 at 8:14 pm

It’s been fourteen months since I wrote that post and in that time frame, what has AT&T done in the CDN market? I think my point is proven.