Amazon, Google clock best cloud uptimes as Azure struggles

Cloud statistics show Amazon and Google with the least downtime in 2014; many lesser-known players and newcomers sported impressive numbers as well

happy king cloud smile
Credit: akakumo via Flickr

If uptime and availability are what matter most in a cloud provider -- for many, they are -- Amazon and Google currently lead the pack. Azure, on the other hand, has a ways to go.

Cloud-service analytics firm CloudHarmony, which harvests statistics about the uptimes and reliability behaviors of dozens of cloud hosts, recently released statistics detailing total availability across 2014 for all the hosts it tracks.

The list includes major names like Amazon and Azure, up-and-comers like SoftLayer, midtier players like Rackspace, and other, smaller providers like Vault Networks and DigitalOcean. Tracked services don't include only cloud VM services like Amazon's EC2, but also CDN, DNS, object storage, and application engines.

Amazon's high scores probably comes as little surprise, given the company's efforts in solving problems of scale from the inside out. Outages on Amazon are typically measured in minutes or seconds, not hours; its single worst EC2 outage of 2014, by CloudHarmony's tally, was a 90-minute interruption in its southeast server region.

amazon cloudsquare CloudHarmony

CloudHarmony's tally for Amazon's cloud service uptime through 2014. Amazon had outages, but they lasted on average seconds or minutes rather than hours.

Google's cloud performance was one of the big positive surprises in the survey. Its App Engine service clocked a remarkable 99.9988 percent uptime, with a single outage of six minutes and change for the entire year. Compute Engine was slightly less robust, racking up 66 outages for a total of 3.28 hours of downtime. But barring a couple of major, 30-minute hiccups in Google's asia-east1 region, most outages were barely more than a minute or two each.

Microsoft's Azure, on the other hand, racked up 103 outages for a whopping 42.94 hours of downtime. Among the worst of the bunch was an outage in November where Microsoft at first didn't acknowledge the problem, then spent a frustrating amount of time delivering updates to customers about an issue that turned out to be human error. The duration of the outage was nothing compared to the lack of communication from Microsoft and the inaccurate information bandied about. (Microsoft reported all was well even when Azure VMs in Europe were still failing.)

Many of the lesser-known players and newcomers sported impressive numbers. Given their smaller stature, that may be a reflection of the lighter load placed on their services. 

IBM's SoftLayer wasn't ranked on the basis of its virtual-machine offerings, but its object storage, CDN, and DNS services all sported 100 percent uptime. A cloud service from Joyent, the folks behind Node.js, racked up 99.9945 percent (1.59 hours of downtime total) -- a better score than GoGrid, DigitalOcean, or RackSpace.

Verizon, which last week drew ire for a planned two-day outage for part of its cloud, was not tracked in the CloudHarmony survey.

From CIO: 8 Free Online Courses to Grow Your Tech Skills
Notice to our Readers
We're now using social media to take your comments and feedback. Learn more about this here.