The power outage that knocked MySpace offline was the second major failure in the past year at the telecom building in Los Angeles where MySpace houses much of its operations. The downtime at the Garland Building (1200 W. 7th) left MySpace users wondering how one of the web's busiest sites could go dark so easily. Adding to the headscratching was the fact that MySpace is a customer of Limelight Networks, a content delivery network that should (in theory) provide distributed caching and storage.
Other tenants at the Garland Building that were affected included DreamHost, which reported on the weekend power outages on its company blogs. The building lost grid power during a series of rolling blackouts that affected the Los Angeles area as California coped with stifling heat and record demand for electricity. The loss of power at Garland was blamed on the failure of a UPS system (uninterruptible power supply), which normally maintains power to equipment while generators start up.
Equinix said that reports that the outage may have occurred in one of its data centers were incorrect. "Equinix’s data centers were operating as normal over the weekend and this morning, providing consistent power to all customers," Margie Backaus, chief business officer of Equinix, reported in the comments section at GigaOm. "MySpace operates in Equinix data centers, but it also hosts operations in other data centers where we understand the incident occurred." MySpace recently announced that it would lease space in the new Equinix center in El Segundo.
Social networking site MySpace.com is offline currently, attributing the downtime to a power outage in its data center.
It was inaccessible for more than 90 minutes, and then displayed a brief message alerting users to the problems: "Hey everyone! There's been a power outage in our data center. we're in the process of fixing it right now, so sit tight. - Tom" (presumably MySpace co-founder Tom Anderson).
With MySpace unavailable, many of the service's users began posting blog items at LiveJournal, another free blogging service popular with U.S. teenagers.
A dynamically updating chart of the web site performance of myspace.com is available. Netcraft offers a web site performance monitoring service that provides detailed uptime charts, along with e-mail alerts when an outage occurs.
Ranking by Failed Requests and Connection time,
June 1st - 30th 2006
iPowerWeb is the most reliable hostinig company site in June, followed closely by Hostway, as budget hosts continue to demonstrate that their networks can compete with those of high-end managed hosting providers.
iPowerWeb's shared hosting accounts start at $7.95 a month for packages that include a free domain and 10 gigs of disk space. Hostway, which ties for second with Above.net this month, offers "SuperPower" shared hosting accounts that include 150 gigs of disk space and 1,500 gigs of data transfer (no, those aren't typos) starting at $9.95 a month.
Leading managed hosting providers continue to turn in strong performance, with Datapipe, Navisite, Rackspace and New York Internet all among the top 10, which included four sites on FreeBSD, three on Linux and two using Windows Server 2003. Of the 50 major hosts we monitor, 34 had no measurable outages in June.
Network problems at Internet backbone provider Level 3 have caused some intermittent connectivity problems today for customers of several U.S. access providers. The issue appears to be related to network integration of Wiltel Communications, which was acquired by Level 3 last fall. By late afternoon, East Coast time, Level 3 was reporting that the situation had been stabilized. Level 3 is one of the largest providers of wholesale dial-up service to ISPs in North America and is the primary provider of Internet connectivity for millions of broadband subscribers through its cable and DSL partners.
Many sites hosted at Fasthosts, one of the UK's largest hosting providers, were unavailable for about 9 hours Sunday. Fastshosts hosts approximately 560,000 hostnames. The outage knocked out Fasthosts' own site, along with customers including Net4Now, which attributed the problems to a fiber cut at Telewest/BlueYonder. The fiber cut also affected cable TV service to about 100,000 UK subscribers who lost their signal just prior to the start of Sunday's World Cup soccer game, in which England defeated Ecuador 1-0.
Net4Now wondered how a fiber cut at a single connectivity provider could affect so many sites, given Fasthosts' statement that "every server enjoys a fully redundant Tier 1 connection so our customers never experience slow-down or service interruptions." The outage can be seen in this chart of the uptime for Fasthosts' main web site:
A dynamically updating chart of Fasthosts' web site performance is available.
Several popular UK-based betting online betting sites have been hit by brief outages this morning, interrupting wagering on the World Cup. Sites operated by BetDirect and UKBetting were offline for periods of between 45 minutes and two hours during the same general time period.
Industry observers say as much as £1 billion ($1.84 billion U.S.) could be wagered on this year's World Cup, five times the betting volume for the 2002 event. It's not clear whether heavy traffic was a factor in today's outages, but that type of betting volume makes downtime quite expensive for online bookmakers. Historically, betting sites have been targeted by DDoS extortionists during major sporting events.
Netcraft offers a web site performance monitoring service that provides detailed uptime charts, along with e-mail alerts when an outage occurs.