A cloud portfolio management provider has done some Monday-morning quarterbacking through cloud hosting customers affected by the recent Xen hypervisor problem and reboot for major cloud players.
If you go by Rightscale's numbers released today -- they surveyed about 449 customers of Amazon Web Services, Rackspace and SoftLayer -- Amazon seems to come out on top as the hosting companies worked to patch and reboot the potential issue announced Oct. 1.
Rightscale reported that of the 349 Amazon Web Services customers responding to Rightscale's Oct. 2-3 survey, 51 percent reported no application downtime as a result of the reboot, and another 21 percent reporting less than five minutes of downtime.
Of the 66 Rackspace Public Cloud users, 27 percent escaped with no downtime. Of the 42 SoftLayer Virtual Server respondents, 26 percent reported zero downtime.
Meanwhile, 5 percent of AWS users reported more than one hour of downtime, while 13 percent for Rackspace and 17 percent for SoftLayer said the same.
(Rightscale also received 74 responses from organizations that used Xen in their internal data centers. Some respondents use multiple clouds, hence the total of more than 449).
Why This Survey?
Reached by CMSWire to comment on RightScale's survey, an AWS media representative referred to this and this blog post. She added less than 10 percent of the Amazon Elastic Compute Cloud (EC2) fleet was impacted.
Kim Weins, vice president of marketing at RightScale, responded to CMSWire's question about its survey methodology, saying the company "periodically surveys a broad mix of cloud users." She cited its annual State of the Cloud report.
In the case of the Xen reboot issue, 55 percent of respondents were in IT/Ops, 38 percent were in development, 4 percent were in business roles and 3 percent were "other." RightScale users represent about one-third of the sample.
Is RightScale competing with any of the Xen reboot companies?
"RightScale is a cloud portfolio management platform that enables customers to deploy, manage and govern applications across public and private clouds, as well as virtualized environments," said Weins, also the author of today's RightScale blog post. "None of the cloud providers are competitors. In fact we integrate with all of the cloud providers included."
Story in the Numbers
What else did cloud customers say?
AWS users benefitted from preparation. Some users (29 percent) relaunched instances ahead of the reboot and moved resources to unaffected instance types (12 percent), according to RightScale. And 43 percent of them used multiple availability zones (AZs).
Rackspace and SoftLayer users were limited to "moving resources between regions" -- 15 percent Rackspace users and 18 percent SoftLayer.
"As a result," Weins reported, "48 percent of Rackspace users and 39 percent of SoftLayer users took no preventative action, as compared to 20 percent of AWS users."
Can We Do Better?
Despite the reboot issues, only 5 percent told RightScale they expect to use the cloud less.
But in response, they do intend to create a plan (37 percent), implement redundant architecture across AZs (35 percent) or regions (27 percent), employ more automation (30 percent) and leverage multiple cloud providers (26 percent).
Could any of these companies done anything differently as far as their response to the Xen issue and subsequent reboots?
"These type of events happen in the datacenter and they happen in the cloud," Weins told CMSWire. "The cloud providers needed to patch servers to address the vulnerability. In either case, a focus on automation makes these type of events much less of an issue for enterprises."
The main thing companies can learn from the Xen issue, Weins added, is to expect issues like this in the cloud, just like they expect them in existing data centers.
"Once the correct architectures, automation and processes are in place," she said, "this kind of event is much more easily handled."
Weins told CMSWire that other major cloud providers, such as Azure and Google, do not rely on Xen. Organizations using Xen in their internal data centers, however, needed to address this vulnerability.
Title image by Whiteaster /Shutterstock.