AT&T’s CDN re-launch
(This is part of a series of “catch-up” posts of announcements that I’ve wanted to comment on but didn’t previously find time to blog about.)
AT&T recently essentially re-launched its CDN — new technology, new branding, new footprint.
AT&T’s existing CDN product, called iCDS, has had limited success in the marketplace. They’ve been a low-cost competitor, but their deal success in the high-volume market has been low — Level 3, for instance, has offered prices just as good or better, on a more featureful, higher-performance service, and with other competitors, notably Akamai and Limelight, willing to compete in the low-cost high-volume market, it’s been difficult for AT&T to compete successfully on price (although they certainly helped the general decline in prices). We’ve also seen them get good pick-up with CDN added to a managed hosting contract — there are plenty of managed hosting customers happy to sign on $1000 or $2000 worth of CDN. (We’ve also seen this with other hosters that casually quote a little bit of CDN along with managed hosting deals; it’s not just an AT&T phenomenon.) We’ve also seen them pitch “hey, you should use us if you want to reach iPhone customers”, but that’s too narrow for most content providers to consider right now.
Previously, AT&T had been insistent on developing all of its CDN technology in-house. AT&T has a long and proud “built here and only here” tradition, especially with AT&T Labs, but it simply hasn’t worked out well for its CDN, especially since anything that AT&T builds in portal technology tends to look like it was built by hard-core geeks for other hard-core geeks — not the slick, user-friendly, Web 2.0 interfaces that you’ll see coming out of many other service providers these days. That made all iCDS things to do with “how customers interface with the CDN to actually get something useful done”, including configuration and analytics, pretty sub-par to the market.
AT&T has now done something that would probably be smart for other carriers to emulate — buying CDN technology rather than developing it in-house. There are now plenty of vendors to choose from — Cisco, Juniper (Ankeena), Alcatel-Lucent (Velocix), Edgecast, 3Crowd, JetStream, etc. — and although these solutions vary wildly in quality and completeness, I’m still bemused by the number of carriers whose engineers are really jonesing to build their own in-house technology. In AT&T’s case, they’ve selected Edgecast’s software solution — a nice feather in the cap for Edgecast, definitely, given the kind of scrutiny that AT&T gives its solutions that are going to be deployed in its network. (Carrier CDNs are very much a hot trend at the moment, although they’re a hot trend relative to the otherwise glacial speed at which carriers do anything.)
AT&T is building out a new footprint of servers running the Edgecast software. They’ll operate both the old and new CDNs for some time — existing iCDS customers will continue to run on the existing iCDS platform and footprint, and new customers will go onto the new platform. That means it’s going to take some time to assess the real performance of the new CDN, as the POPs are being rolled out gradually. The new footprint will be similar, but not identical to, the old fotprint.
However, I don’t think the launch of a new AT&T CDN is anywhere near as significant for the market as AT&T’s continued success in reselling Cotendo. The AT&T CDN itself is simply part of the already-commoditized market for high-volume delivery — the re-launch will likely return them to real competitiveness, but doesn’t change any fundamental market dynamics.