One of the lessons learned during the Amazon Web Services (AWS) outage this week is that you need more than one.
Relying on one of anything creates a certain amount of risk. If you have one car and it breaks down, you're stuck.
Tuesday's outage shows what can happen when you rely on only one region of any cloud service. Looking back on the outage, we can see that a number of large organizations haven't fully engineered their applications or services to be fault tolerant, resilient, and flexible, or to have a disaster recovery plan in place.
The outage affected C2's CMS partner, Episerver, at its annual Ascend conference in Las Vegas. During the outage, attendees were not able to post images, stored in S3, to the in-conference mobile app.
There are a number of reasons that these organizations may have chosen the solution impacted by Tuesday's outage. These range from cost to complexity, and some organizations don't require 100% uptime and availability.
AWS S3 has a built-in feature called Cross-Region Replication, which automatically copies data stored in an S3 region to another or additional regions. This provides redundancy should disaster strike a region where your applications or solutions retrieve data.
Adding to Cross-Region Replication, the use of a content delivery network (CDN), like AWS Cloud Front or Akamai, provides additional resiliency and ensures the public doesn't realize an application was impacted by an outage.
And by disaster, we don't necessarily mean earthquakes, hurricanes, and other natural disasters. These features act as insurance against human error, too, which was the case Tuesday, when a typo was reported to be the culprit for the outage.
I'm glad I have two legs, two arms, two cars, etc. If you'd like to learn more about how C2's managed services team can help architect and build a highly resilient, fault-tolerant cloud environment for your applications or services, feel free to contact me.