Last week Amazon Web Services suffered a high-profile outage. Early Thursday morning at around 1am, the Elastic Block Storage (EBS) service failed in one of their Virginia datacenters, bringing down several large websites and hundreds of smaller ones. The failure in this data center caused the load to cascade to other AWS datacenters, causing additional problems. These issues took several days to fully recover, raising many to question the viability of hosting applications in "The Cloud" instead of managing your own servers.
At Instructure I manage a team of experienced software and operations engineers that have been building large scale cloud systems for the better part of a decade (yes, "the cloud" is not actually that new) and we are experts so our customers don't have to be. Instructure didn't experience any problems related to the AWS outage, even though we do have servers hosted on AWS in the Virginia datacenters and we do utilize the EBS service. Our system is spread across several datacenters to maximize availability, even in a disaster scenario like last week.
Does this mean that Instructure will never have downtime? No, of course not. No one can guarantee 100% uptime - but we can assure you Instructure is based on a robust global architecture that we obsess over 24 hours a day, 365.242199 days per year, so you can spend your energy obsessing over your passion for education.
Director of Engineering and Operations