(WEB HOST INDUSTRY REVIEW) — A data center outage that affected some of Rackspace’s (www.rackspace.com) cloud hosting customers Friday led to widespread criticism of the company from some of its customers that are popular bloggers or users of social media sites.
According to a post on the Rackspace blog, the outage was caused by a networking issue, and affected its Dallas Fort Worth data center, and lasted from 3:37 p.m. until 4:17 p.m. central time.
“The issues resulted from a problem with a router used for peering and backbone connectivity located outside the data center at a peering facility, which handles approximately 20% of Rackspace’s Dallas traffic,” said the post.” The problems stemmed from a configuration and testing procedure made at our new Chicago data center, creating a routing loop between the Chicago and Dallas data centers.”
For some customers of Rackspace, the issue is less about the single, and relatively brief, period of downtime, and more about the company’s track record over the past year, which has seen several high profile outages, including one in June.
Perhaps more significantly, the nature of many of Rackspace’s customers – many influential tech websites and popular bloggers among them – means that the company is under an almost unnatural magnifying glass that amplifies the impact of this kind of service interruption.
Among the sites impacted by Friday’s outage were popular technology blog TechCrunch, sites hosted by Tumblr, and sites hosted by Rackspace reseller and popular online community LaughingSquid – a sort of perfect storm of people capable of getting the word out, even if the word is negative.
In a Friday post on TechCrunch, a glib and demonstrably unimpressed MG Siegler wrote, “this is another black eye for the company, though they are generally responsive with other issues we’ve had throughout our time with them. But until they can prove to be more reliable, we’ve decided to get a backup version of TechCrunch up and running at another datacenter, for when someone inevitably trips over a power cord at the Dallas Rackspace center again.”
A weekend post on the CIO Weblog by Scott Wilson addressed the fact that courting “influential bloggers” is, in a sense, inviting this kind of backlash should anything go wrong. He also pointed out that Rackspace’s cloud offerings seem to, counter intuitively, be run entirely out of the single data center in DFW. He suggest that geographic redundancy ought to be a sort of no-brainer inclusion on the list of criteria for what constitutes a cloud.
