Service Instability Underscores Serious Cloud Issues – and the Need for Better SLAs

A. DeCarlo

Summary Bullets:

• An unfortunate series of big impact cloud outages including Windows Azure, Salesforce.com, Amazon Web Services, and Twitter have users fuming and organizations rethinking their on-demand strategies
• Are cloud providers doing enough to address the underlying issues and reassure enterprises?  The answer, at least for now, seems to be no

Summer malaise has hit the cloud in a big way with a series of service interruptions knocking some of the most popular services offline temporarily and setting some corporate users into a tailspin. Though the root causes of the failures may differ, providers often issue ineffectual mea culpas, often citing avoidable issues like Twitter’s “double system failure” or Salesforce.com’s power failure rather than more unpredictable causes like a natural disaster or an unanticipated massive influx of traffic flooding their service.  The result leaves already leery enterprises even more on edge about making the shift to the cloud anytime soon.

Though most providers have tried to explain the source of the failure, few are doing enough to regain the confidence of their users.  The first step to reassuring corporate clients, beyond the obvious need to restore service as swiftly as possible, is to clearly acknowledge the outage and, as much as possible, identify the source and explain not just how it is being fixed but also what steps are being taken to prevent future issues.

Cloud providers should also look at longer term ways to inspire customer confidence, starting with more ironclad, proactive service level agreements that extend beyond flimsy uptime agreements to incorporate elements that focus on consistency.  Though providers may consider being more aggressive with credits in the event of an outage, what customers really want isn’t money off their bill but rather the assurance that their service will be accessible, reliable, and high performing.

Has your business suffered as the result of a cloud outage?  Were you satisfied with how your provider addressed the issue?