By John Agsalud
Users of Microsoft’s Office 365 email service, commonly known as Exchange Online, experienced an outage of about three hours a couple of weeks ago. Most of these users had their business processes completely short-circuited due to ever-increasing reliance on email. What, then, should businesses and government agencies do to alleviate issues when (not if) something like this happens again?
First, organizations need to identify the critical components of their business infrastructure and budget for appropriate fallback and/or redundant solutions. In an ironic twist, the recent outage provides an excellent measure of a business unit’s dependence on email. After all, the issue occurred during prime-time working hours (reports started coming in as early as 2:30 pm local time, and Microsoft didn’t call the issue resolved until after 5 pm).
How much did that outage really cost you, both tangibly and intangibly? This is probably a good time to assess that, now that some time has passed to allow emotions to cool.
Wait a minute, you say, didn’t I go with cloud-based email so I wouldn’t have to worry about a fallback plan? Isn’t that supposed to be the provider’s problem? Well, yes, but unfortunately, no cloud-based solution is perfect, because humans are always involved. In fact, in this case, Microsoft blamed human error (implementing a software update that had not been fully tested) as the root cause of the outage.
So what kind of redundancy can you implement for cloud-based email? In many cases a handful of email accounts from an alternate provider would help. If your primary provider is Microsoft, for example, sign up for Google accounts. While not ideal, emergency business processes certainly can be engineered to help you get through an outage. Test your processes and figure out what you need and how to use them in case of emergency.
What about your Internet connection? While not relevant to this particular outage, we see many organizations that have only a single connection to the cloud. With an increasing dependence on cloud-based solutions, more bandwidth is needed, and a backup connection is also becoming more and more common.
When acquiring a backup Internet connection, be sure to inquire about something known as path diversity. Somewhat unknown to many, Provider A often just resells circuits that are actually delivered by Provider B. Or, Provider A uses the same wires as Provider B. Path diversity also takes into account the conduits and routes a circuit might take. Even if different wires are used, if they end up in the same underground conduit or on the same telephone pole, one fender bender could take down multiple providers.
Finally, users of Office 365 should be aware of their Service Level Agreement (SLA). The standard SLA for Office 365 says that if the service is not up for 99.9 percent of any one month, users are entitled to a 25 percent credit of their monthly fee. This translates into about 45 minutes for a 31-day month, which clearly and by Microsoft’s own admission was exceeded. So be sure to file a claim for your credit if you think it’s worth it.
