The Cost of Cloud Downtime Starts to Become Clear
A new report takes a look at cloud provider outages over the past few years and makes some remarkable conclusions about the cost associated with these outages. However, some issues with the report's methodology should give IT managers pause before taking the results to heart and altering their cloud strategy.
Cloud Downtime Report
The report was issued by the International Working Group on Cloud Computing Resiliency (IWGCR), which was founded a few months ago by Telecom ParisTech and Paris 13 University. The group found that since 2007, the 13 largest cloud providers have recorded over 500 hours of downtime to the tune of an economic impact of around $70 million dollars.
As noted by IT Pro, since many providers don't offer information about their own outages after the fact, the group had to scour press releases from a number of industry websites to collect the information about the outages. They then used hourly costs accepted by the industry to attach a dollar figure to the loss due to the cloud downtime.
The report places the overall cloud uptime at around 99.9 percent, which, as Wired points out, is a far cry from the 99.999 percent uptime that is required for mission critical applications. The report at the IWGCR website shows a 7.5-hour average amount of downtime per year per cloud client. When compared to the 15 minutes of downtime allowed for electricity in modern capital, it shows just how far the cloud has to go.
All reports like this need to be taken with a grain of salt, and this one more than most. The IWGCR acknowledges some of the report's failings within the text itself, noting that their method of data gathering is far from exhaustive. Additionally, their data is not based on the number of users, and their calculations for cost are imprecise. Beyond what the group acknowledged, there are other issues with the data. The three-day BlackBerry outage was included in this data, although few would argue that it should be considered a black mark on the cloud computing industry. The massive Amazon Web Services outage of 2011 was also not included, and it is hard to believe that there weren't press releases about that headline-grabbing downtime.
Finally, few would consider that the cloud of 2007, if it was even really a cloud back then, resembles the modern cloud in any form. The most significant outages in the report happened prior to 2010, and if the BlackBerry outage were removed from the data, the report only showed 20 hours of downtime after 2009.
Managing for Eventual Downtime
Regardless of the conclusions you could draw from the cloud downtime report's data, the chance of a cloud outage is always present. But that doesn't mean that IT managers should give outages too much weight when deciding on a cloud solution. Midsize businesses are ideally suited to take advantage of the extra power and scalability that the cloud can offer, and while downtime can be dangerous for the bottom line, that risk will be present no matter what kind of technological solution the business uses. Neither non-cloud remote hosting nor internal infrastructure are immune from outages, and many suffer from downtime far more frequently than today's cloud industry leaders.
Still, planning for eventual cloud outages is always a good idea. One of the best things IT managers can do is to ensure that their software is restartable, meaning that if the servers go down for a brief amount of time and then come back up, the application will automatically restart. IT managers can also look at cloud providers that offer clusters, allowing clients to have backup versions of the application at the ready in case one cluster goes down. Clusters are supposed to be completely separate from each other, although it turned out that Amazon shared storage across clusters, resulting in an embarrassing outage.
At this point, any midsize firm that doesn't have a ton of capital already invested in appliances needs to give the cloud some serious consideration. Cloud providers continue to attract clients precisely because what they have to offer is so remarkable. Any technological solution will have some downtime, so IT professionals should not let reports of minor downtime deter them from adopting a solution that is changing the industry.
This post was written as part of the IBM for Midsize Business program, which provides midsize businesses with the tools, expertise and solutions they need to become engines of a smarter planet. Like us on Facebook. Follow us on Twitter.