Outages will happen. That makes your job knowing what to expect when they do.
Last week's Microsoft Skype outage, which lasted about 15 hours, can give enterprises that are adopting cloud services some idea of the risk they might be taking on -- and can serve as an important lesson for everyone.
As widely reported, some Skype users experienced technical issues on that knocked the service offline on Monday, Sept. 21. In the official Microsoft "Big Blog" post on the outage, Gurdeep Pall, corporate vice president for Skype and Skype for Business at the company, wrote:
We are extremely sorry for any inconvenience caused to our users, and appreciate your patience while we addressed the issue.
He noted, however, that Skype for Business was not impacted.
Since Skype use comes without a service-level agreement (SLA), an apology is all users get. And, really, that's all they can expect of cloud technology on the cheap.
As you consider cloud services, begin by determining whether you can get by with a a free service or whether a paid model better fits your needs for reliability (among other factors). And for paid cloud services, make sure you understand the outage clauses within your SLAs. How long does the service need to be offline before the provider issues a credit, and would that credit effectively cover the cost of the business disruption? Have you assessed those potential costs? You should.
Is your provider keeping you in the dark or will it reveal the scope of the outage? Not knowing the nature of the problem or when to expect resolution makes people crazy.
With cloud services, the fixes or resolutions aren't in your control -- and that can be frustrating for some as well. You can have a rock-solid SLA, but it won't provide any relief -- regardless of how well your contract is written or what language you use, you remain at the mercy of the provider's ability to fix the technology and restore service. When this doesn't happen in a timely fashion, then you need to know how to respond.
Remove the words "Skype" and "Microsoft" in this article and replace them with the name of any cloud service and provider. What's in place in your cloud, and how will your services be affected by an outage? Unfortunately, many times companies don't address these key concerns because their cloud services go untested. Just because a contract says a backup is in place, that doesn't mean you don't need to test it. How else will you know that the contingency service to the primary cloud actually works?
Perhaps expectations aren't always what they should be, but enterprises need to be aware of what they actually are. Because, in the grand order of the cloud, "cloud down" is not a happy thing to face -- not on a Monday and definitely not on a Friday afternoon.