2 Comments

Summary:

For anyone needing a reminder that no computing cloud is perfect — Microsoft’s Windows Azure cloud went down in Europe on Thursday. Few details are available, but Microsoft says it’s on the case.

4941351962_cc250d9a0b_z

Just in time for the Olympics, there’s a Windows Azure outage in Europe.

A flag posted at 11:09 UTC  on the Windows Azure dashboard: 

“We are experiencing an availability issue in the West Europe sub-region, which impacts access to hosted services in this region. We are actively investigating this issue and working to resolve it as soon as possible. Further updates will be published to keep you apprised of the situation. We apologize for any inconvenience this causes our customers.”

According to an update posted one hour later, the troubleshooting and data gathering continued.

This is just the latest reminder that while the cloud offers customers great scale and potential cost savings, not even the largest, most well financed public clouds are immune from downtime. Windows Azure experienced a widely publicized “Leap Day” outage in early March. Amazon, the largest public cloud provider, was hit by several outages at its U.S. east data center over the past few months.

Update: At 12:46 p.m. EDT, I got this update from Microsoft:

The recent Windows Azure service interruption in the Western Europe sub-region has been resolved. We apologize for any inconvenience this outage may have caused our customers.  The duration of the service interruption was approximately 2.5 hours and was resolved at 6:33 AM PDT.  Customers who have questions regarding this incident are encouraged to contact Customer Service and Support.

Stay tuned for updates.

Photo courtesy of  Flickr user Stew Dean

 

  1. What is the cost of downtime? Events like this seem to suggest that it is too risky to use the public cloud for anything that is critical for day to day operations. To me this seems like the exact opposite of using the network traditionally for BCDR. In the BCDR case, the WAN is acting as connectivity to standby resources, and outages here don’t halt business.

    Share
    1. Timothy Weaver Thursday, July 26, 2012

      I’m unfamiliar with Azure’s architecture, but in the AWS situation you need to design your solution differently if you want to achieve higher service levels. It’s definitely doable, and there are plenty of businesses running critical day-to-day solutions on AWS. For example, designing your application to depend on a single Availability Zone (essentially data center) is going to cause a problem if that AZ has problems. But AWS has multiple regions and multiple AZs in each region. You need to architect to split across those environments to achieve the highest availability. But it’s doable and there’s nothing inherent about IaaS preventing it. If you can design internally for something, you can design for it with IaaS too.

      Share

Comments have been disabled for this post