In the past few weeks there have been major outages by a high profile cloud infrastructure provider and a large SaaS vendor, both of whom have had a reasonably good track records. Unfortunately, much of the attention in the press and even by some analysts has been devoted into what went wrong versus providing advice to customers on what to do to protect themselves. Explaining the circumstances that led to an outage maybe technically interesting and provide some insight into overall risk of a provider but how does this help a customer or a new customer of the cloud service. I think the most interesting comment by someone was that one of the affected cloud providers was down for several hours but was still within their uptime commitment for the year. Huh? That is supposed to be the good news.
There is one very simple lesson customers needs to take away from these high profile outages, a service is only as good as a cloud provider is willing to back up with an SLA. As in the financial markets, past performance does not dictate future performance. Just because a vendor has had a good track record for uptime does not mean they will have good uptime in the future. Gartner has published research on best practice SLAs and will go a long way to helping customers to protect their interests when contracting for cloud related services. Don’t take a vendors word, or track record get commitment for performance.