The Effects Of The Amazon Web Services Outages

The Effects Of The Amazon Web Services Outages

There has been no major cloud computing outage in 2012. However, non-believers had a field day when Amazon Web Services customers experienced some service disruptions on June 14. Although AWS went offline for a few hours only, the downtime experience did have an impact on customers’ businesses. According to reports, a power outage struck Amazon’s Northern Virginia datacenter. Internet sites and startups such as Heroku, Quora, Pinterest, and Parse were affected. Heroku filed an incident report. As per Amazon’s Service Health Dashboard, the problem was reported before 9 p.m. Pacific Time. The power outage affected AWS Elastic Beanstalk, Amazon Relational Database Service, and Amazon Elastic Compute Cloud, which are all housed at its datacenter in North Virginia. The services were fully restored in a few hours.

The outage was clearly not IT-related, but Amazon did receive a lot of flak for the service disruption. A large number of comments have been posted online from people from all walks life, even from the company’s competitors and non-clients. In a nutshell, the lesson cloud computing services users must learn is that they must spread their workloads across various parts of Amazon Web Services in order to prevent being hugely affected if an individual AWS region experiences service disruption. In most cases, such service disruption is often blown out of proportion. Customers who are having doubts regarding the credibility of cloud computing service providers will likely stay out of the cloud and still prefer to buy more expensive IT products hosted on-premise.

In April 2011, the North Virginia datacenter suffered a similar service outage which affected GroupMe, Quora, Reddit, HootSuite, Twitter, and Foursquare. Amazon reported that the outage was a “re-mirroring storm”. At that time, only the relational database service and the Amazon EC2 at the North Virginia center were affected, but it did affect many customers’ services and websites. What is surprising is that even though AWS experienced some outages, the company is continuously growing. Amazon reported that its S3 storage has more than a trillion objects, which is 143 times the population of the world.

According to a recent report by the International Working Group on Cloud Computing Resiliency, a minimum of 10 hours are lost because of service disruptions yearly. The group found out that since 2007 a minimum of five hundred hours has been lost as reported by the thirteen biggest cloud computing service providers, which can be translated in economic terms to be worth a minimum of $70 million. In that same report, the group claims that a cloud computing service is usually down for an average of 7.5 hours each year, although an electric power service outage is pegged at a low 15 minutes yearly.

The group gathered the data from various sources such as Twitter, Amazon, Google, Paypal, Yahoo, Microsoft, Facebook, and others. It is estimated that the travel service provider Amadeus loses $89,000 per hour during any cloud computing outage, while Paypal loses around $225,000 per hour. There is no known data for the number of people affected by a cloud computing service outage.

By Florence de Borja

3 Responses to The Effects Of The Amazon Web Services Outages

  1. Businesses need to think how they want to deal with outage risks. At first companies kept mission critical processes in house. They can also adopt a multi-cloud operating model. Spreading their resources over several sites and sometimes several providers, ensure that if a power failure brings one data center, other still can work.

  2. @danielsteeves which #cloud to use shows the need for a strong risk management approach. Compliance, security, availability etc.

  3. @sarojkar
    The effects of cloud computing in perspective of a vendor are virtually inevitable. AWS are pioneers of cloud and certainly might have the most powerful cloud infrastructure but still they face such problems. For an end-user/business perspective, as Christian pointed out, we need to have a strong risk management and mitigation policy in place. Perhaps, an expensive but feasible route would be to have another cloud vendor as your DR/Backup provider which can provide the most essential/required infrastructure in-case of such down-times.
    If we do simple math, Paypal loses about $225,000/Hour as Florence pointed out and potentially it can lose a million$ if they are down just for 4hours annually. But having a third-party DR/Backup solution might cost considerably less than this if computed annually.  

The Lighter Side Of The Cloud – The Apple Watch

Contributor Spotlight

Steve Prentice
Jennifer Klostermann

CloudTweaks is recognized as one of the leading influencers in cloud computing, infosec, big data and the internet of things (IoT) information. Our goal is to continue to build our growing information portal by providing the best in-depth articles, interviews, event listings, whitepapers, infographics and much more.

CloudTweaks Comic Library