Cloud over the Cloud?

If you haven’t followed the development of cloud computing over the last few years then you may not know what AWS stands for, or that Amazon is one of the world’s largest cloud computing platform providers in the world.

Amazon Web Services (AWS) provides the backbone for many popular website and online services that many have come to rely upon.  Last week’s AWS outage caused service disruptions to many high profile websites creating a firestorm of criticism throughout media and cloud computing circles. Some, even calling into question the viability of the cloud.  I saw one title that summed up the reaction perfectly “The Sky is Falling, AWS is Down”.

First, let me assure everyone who reads this article that the sky in not falling and without going into too much techno-jargon, AWS wasn’t down, they had isolated outages to one availability zone.  I’m not trying to minimize the disruption, which was massive. However, proper systems architecture would have gone a long way to limit the impact to performance issues and may have even eliminated downtime altogether. 

Since the advent of the cloud computing concept, few principles have been leveraged more than reliability. Reliability, scalability and economics to scale are three of the core concepts behind the success of cloud computing platforms.

Given the very nature of the “Clouds” reliability many seem to have opted not to incur the added expense of designing and deploying truly redundant systems. While this cost is significantly less than a truly redundant self-hosted solution there is added expense over a system deployed in a single availability zone.

The cloud is still and has been since its advent the most reliable, scalable and cost effective way to acquire and manage online system resources, without question. The moral of the AWS outage, if you need to ensure 100% availability, the cloud doesn’t elevated the need to implement a system design that ensures you will meet that mandate.