Take advantage of our commitment to Fanatical Service, enhance your Peace of Mind and get Fully Managed™.
by Chris Day @ Fully Managed™ Over the past couple of weeks, we've seen a couple of significant system outages from some of the world's largest and most well-respected cloud providers: Amazon's EC2 infrastructure outage on April 21st, which lasted a full three days - and more recently, the May 10th Microsoft Online Services outage which carried on for two days. Many are now second guessing whether or not they should still be moving to the cloud in light of these recent service outages. My answer? Heck yes.
Many people have a distorted view of what downtime (the inverse of uptime) is all about. A provider can achieve 99% uptime by ensuring that their service is available for all but 86.4 hours or 3.6 days in a calendar year. By my calculation, Amazon is currently tracking below 99% based on this recent outage, a far cry from their EC2 SLA target of 99.95% (4.32 hours or 0.18 days of downtime per year). SLA is basically a metric designed to make people feel better when assigning risk to a provider. In reality, when you consider what it feels like to incur a 2 or 3-day downtime incident, SLA paperwork offers little comfort.
From what I've seen thus far, achieving 99.95%, 99.99%, or any other similar incredible uptime number is simply not yet possible given the complex engine that drives these public cloud infrastructures. Everyone is still learning, and improving continuously, but we're not there yet. Can anyone prove me wrong? This is why we based our SLA as a Managed Services Provider (MSP) on what we can directly affect - for example, response time.
Consider your next best alternative - either hosting with a 3rd party provider or building out your own infrastructure. Do you think that either of these options will be more reliable or resilient than what Microsoft, Google or Amazon could build for you for a price that meets your budget? Dig deep and I think you'll find the answer.
