Amazon's public cloud outage last week wasn't the worst in the company's history, but that and other recent cloud services interruptions give IT professionals who aren't ready to make a commitment to the cloud reason to continue their resistance.
In the meantime, cloud customers need to take more responsibility for their own uptime, said one long-time Amazon Web Services (AWS) customer."Downtime is lost revenue, and every level of the company needs to take responsibility for emergency preparedness," said Colin Dean, president of the Pittsburgh LAN Coalition, a group that organizes video gaming events."Any time a company decides to use a cloud service, they absolutely must assess the risk an outage poses to their business," Dean said. Although he was not affected by the outage, Dean's critical systems are set up to fall back on local disks or to another cloud service.Last week's outage took down a single Availability Zone in Amazon's data center in northern Virginia, which supports the US-EAST-1 region. However, the outage, which was triggered by a power failure that the company did not elaborate on, affected a disproportionate number of customers because that data center is AWS's largest and its oldest.A list of customers who were without service for several hours late last Thursday and into the early morning hours of Friday was not available. However, Heroku, a division of Salesforce.com, was affected. Heroku, which is one of the largest Platform as a Service (PaaS) providers, runs on AWS.A Heroku official said the company is still investigating the outage but had no further comment. A more severe AWS outage in April 2011 that went on for days also hit the PaaS provider. That may have helped inoculate some customers against heightened paranoia caused by this latest outage, said one analyst."This is a rehash of earlier outages in terms of the fallout … it's loud and sometimes histrionic, but ultimately, [it’s] not going to make a dent in either Amazon's reputation or their customer uptake," said Carl Brooks, analyst for infrastructure and cloud computing at Tier1 Research, a division of New York City-based 451 Research LLC.Still, there are IT factions that see these outages as an opportunity to retake lost territory in the data center."Outages are going to happen ... that said, it will serve to reinforce the stereotypes of cloud computing among the irredentist enterprise faction that is feeling threatened by the trends toward external IT services," Brooks added.In fact, failures like those at AWS, as well as Microsoft's Windows Azure outage in February, continue to provide cover for some IT managers opposed to risking processing and data to the cloud.For IT departments looking for excuses to delay public cloud adoption, this incident might reinforce their argument, said one New York-based cloud computing consultant with extensive experience with AWS.