AWS, Google Public IaaS Clouds were Most Reliable in 2014
Which cloud services providers (CSPs) offered the most uptime in 2014? Cloud services performance analyst CloudHarmony found that the Amazon Web Services (AWS) Elastic Compute Cloud (EC2) and Google Cloud Platform were two of the top-performing public infrastructure-as-a-service (IaaS) clouds in 2014.
While downtime was a problem for some CSPs last year, both AWS EC2 and the Google Cloud Platform maintained the best uptime percentages, CloudHarmony said.
AWS EC2 recorded 2.41 hours of downtime across 20 outages in 2014, which represented 99.9974 percent uptime, according to CloudHarmony.
Also, CloudHarmony found that the Google Cloud Platform storage service experienced 14 minutes of downtime last year for a 99.9996 uptime percentage.
“The more established players are fine-tuning their systems and becoming quite stable,” CloudHarmony CEO Jason Read told Network World. “AWS has been providing cloud services longer than anyone in the market and Google uses its existing infrastructure for its cloud, so it too has a long track record of managing a reliable distributed system.”
Other CloudHarmony findings included:
- The Google Compute Engine component of the Google Cloud Platform experienced 72 outages for 4.42 hours of downtime in 2014.
- Microsoft Azure experienced 92 outages totaling 39.77 hours of downtime in 2014, and its storage platform had 141 outages totaling 10.97 hours of downtime. (Last November’s Azure outage may have affected these results.)
- The AWS Simple Storage Service (S3) cloud storage platform had 23 outages and 2.69 hours of downtime in 2014. About 10 percent of AWS EC2 instances had to be rebooted in 2014 due to a Xen hypervisor issue that was discovered last September as well.
Donnie Berkholz, a senior analyst at industry analyst firm RedMonk, said he believes CSPs are getting better at preventing outages, but CloudHarmony’s data also indicated that some cloud computing buyers are ignoring these outages altogether.
In fact, CloudHarmony found that Azure, CenturyLink (CTL) and Digital Ocean were popular cloud services despite higher-than-average downtime. (CenturyLink experienced about 26 hours of downtime in 2014, while Digital Ocean had roughly 16 hours of downtime.)
“Outage frequency, within a certain range, isn’t a blocker on adoption of an otherwise compelling cloud provider,” Berkholz said. “The question [then] isn’t which provider is best – but what is the limit of what customers find acceptable?”
CloudHarmony monitors more than 48 CSPs for downtime. This cloud services performance analyst maintains web servers in these vendors’ clouds and tracks when their services are unavailable.
CloudHarmony’s full data set is available here.
Share your thoughts about this story in the Comments section below, via Twitter @dkobialka or email me at dan.kobialka@penton.com.
