Microsoft’s Office 365 email service was down for several hours last week, the latest big-name cloud service provider to suffer problems. While the outage may be something of a black eye for Microsoft, it could well be fodder that helps managed service providers (MSPs) start conversations with customers that lead to additional sales.
Microsoft’s outage is just the latest reminder that cloud services are susceptible to outages, just as any IT service is. In fact, Microsoft’s service level agreement (SLA) for Office 365 specifies only 99.9% uptime. That translates to more than 43 minutes of down time per month, and nearly 9 hours per year. That’s what customers should be planning for.
In many cases, they don’t. Customers tend to buy into the promise of cloud – good services available at reasonable prices from wherever users may be – without adequately considering the potential downside, that being what happens when the service is suddenly unavailable.
Any good IT service comes with a backup plan. When customers host their own email servers in-house, chances are they have a backup server in place should the primary fail. Similarly, most build their local- and wide-area networks with some degree of redundancy built in.
MSPs would do well to remind customers that cloud services are no different; you still need a plan B. That’s especially true when we’re talking about such fundamental IT services as email, perhaps one of the most critical applications in many organizations, as well as unified communications, including voice and video. User tolerance for downtime of such applications is extremely low because any outage impacts the company’s reputation and business operations; it may even cost money.
With respect to cloud, I’d say a plan B involves protecting against two primary sources of failure: the WAN connection to the cloud provider and, as in the Microsoft case, the possibility that the service itself will go down.
In terms of the WAN connection, customers would do well to have two different connections, whether it’s plain vanilla Internet service from separate providers or perhaps an MPLS link from one provider with a plain vanilla backup.
Beyond that, customers also need to think about power considerations. The WAN link is only as strong as the router or switch to which it connects. Should power to that device go down, there goes the cloud service. So MSPs should be sure customers have those routers and switches protected by a UPS with sufficient run time. That time will vary dramatically depending on whether there’s a generator on-site to provide backup power or if the device will be relying solely on the UPS until power comes back. That’s a good discussion to have and may well lead to the sale of additional UPS equipment.
Another opportunity is for MSPs to provide a backup hosting service, either from their own facilities or through a third party. Such a service will protect customers when their provider’s infrastructure fails, as in the Microsoft case. Should that happen, customers could fail over to another instance of their cloud service hosted by the backup provider.
Cloud promises to be all things to all people, but it’s not that simple. Yes, it offers generally good quality service at a reasonable price. But it’s not perfect, as the 99.9% SLA indicates.
It’s a bit like the Mick Jagger corollary (queue the choir): “You can’t always get what you want but if you try sometimes you just might find, you get what you need.” Cloud services provide most of what users want, but MSPs can add to the puzzle and provide customers with what they really need – sound backup for those cloud services.