AT&T Builds Out CDN, Prepares to Push Into the Market

Last December, at AT&T's (T) analyst day, the presentation included a few slides about its content delivery build out and capacity planning in 2008 to handle web acceleration, software downloads and streaming based services.

Since December, AT&T has been busy working on the build out and expects to spend between $70-$80 million on infrastructure this year. By the end of 2008, AT&T is aiming to have 400 Gbps of capacity online, for all its content delivery services, which would increase its capacity by 4x what the company has now. When completed, its content delivery services will be delivered from 32 nodes in 7 countries, will be Adobe Flash Certified by year's end and will be supporting live and on-demand delivery for all the major formats.

Currently, some customers of AT&T's are still having their content delivered via Akamai (AKAM), who AT&T has been re-selling and using as one of its partners for some time. But moving forward, AT&T expects to deliver more content across its own network and rely less on partners for delivery. AT&T has been busy training its direct sales force and re-sellers to sell its CDN services and in the third quarter, AT&T expects to aggressively push into the market.

While AT&T won't have some of the additional CDN services in the content eco-system like content management, transcoding, DRM etc... like most CDNs, it will probably partner with others in the industry who provide these services. Its content delivery services already support some additional functionality like authentication, pulling content from customers origin storage and reporting via its customer portal. While AT&T will not say how many customers it has for its CDN services today, or how much revenue it wants to generate from CDN services in 2008, it has listed Forbes.com, AccuWeather.com and the U.S. Golf Association [USGA] as current customers.

While many analysts who cover Akamai were worried when AT&T talked about its CDN plans during the company's analyst day, AT&T still has a lot of work to do in order to become a major player in the content delivery industry. The company does have some advantages going for it, most notable of which is that it is not a startup and not relying on content delivery services alone for revenue. AT&T won't go out of business in 18 months when the VC money dries up, like some of the other CDNs will, and the company has an enormous marketing budget, re-seller channel and plenty of R&D resources. That's not to say those advantages will guarantee AT&T success, as we saw Qwest (Q), MCI, and other telcos in the market fail with these same advantages years ago. But with Level 3 (LVLT) now becoming a major player in the CDN market and AT&T making a bigger push, it's only a matter of time before the telcos once again try to dominate this market.

Some will say that since AT&T, Level 3 and other telcos own the network, that gives them a competitive advantage over CDNs who don't own the pipe. Others say that owning the pipe is too expensive, requires too much capex and does not allow the telco to deliver traffic from multiple "best of breed" networks. At this stage, the verdict is still out on who is right, but one thing is for sure. The telcos are entering the content delivery market and things are going to get very interesting in 2009 when outsourced CDN services for video alone become a billion dollar market in the U.S.

Comments

[User 192430:]

As I asserted when the first piece about AT&T getting more heavily into the CDN space:

Assuming that there is some advantage in having servers in many networks as Akamai does, T will have difficulty in achieving this since competing networks will almost certainly not cooperate.

I would argue that a strong possibility exists of a more extensive partnership between Akamai and T to achieve T's goal.

2008 May 13 04:47 AM

[WetWarrior:]

Mr. Rayburn, I'm still trying to get a handle on the technology overlays governing the telco business model. I follow (and own) SOAP and their (relative) successes in fielding a multi-vendor ethernet provider controller network.

Given SOAPs intimate exposure to T and the fact that SOAP actually built most of T's servers, can the implementation of the Soapstone control plane give T a competitive advantage? Is the technology compatible and/or accretive?

Thanks for a great article.

2008 May 13 10:22 AM

[Dan Rayburn:]

Sorry, I have no insight into SOAP or any expertise in that matter.

2008 May 13 11:56 AM

[User 134567:]

Dan, Level 3 and AT&T already have some sort of working relationship. I'm sure you know the details of this. If AT&T adds CDN services, will they become a competitor with Level 3 or will they continue to work together in some fashion? You may not know the answer to this, but what do you suspect?

2008 May 14 02:22 AM

[Dan Rayburn:]

AT&T said that over time, they will rely less on partners they work with now for delivery and do more of it themselves. That does not happen over night, but slowly more and more delivery will be done by AT&T as they turn up the capacity of the network.

2008 May 14 10:39 AM

[FreeRange:]

So what are the prospects for Level 3 - they have seen good movement recently.

2008 May 16 11:51 AM