In the aftermath of the EC2 outage there has been plenty of blame and bad press about Amazon, but how much of the blame really has to do with them and how much should be placed on the customers for their lack of:
- Defining application requirements
- Using the correct architecture
- Having the proper Service Level Agreement (SLA) in place.
The Amazon customers who experienced an outage had the wrong architecture and the SLA proves it. Amazon claims they didn’t miss their SLA’s yet numerous EC2 customer applications took 24-48 hour outages. Why didn’t someone in product management, system architecture or at least the CTO of the affected companies realize there was no safety net and their application was vulnerable to downtime?
There are countless examples in our everyday lives where we overlook little imperfections. I don’t know if anyone else is like me but occasionally I expect things to go wrong. For example, at least once a day I’m in the middle of a conversation and my cell drops the call.
The EC2 outage was not a little imperfection but something that should have been planned for by the companies who were affected. EC2 is engineered to provide five 9’s uptime, provided the customer uses multiple EC2 regional or geographic zones, which the affected customers did not.
There is no technical mystery or magic to Cloud Computing, it is just a utility model offering, something Ma Bell perfected years ago. The same rules and common sense governing a company’s own infrastructure still apply.
At the company I work for, The Evaluator Group, we consult with large enterprise organizations everyday on developing these business critical components. The core offering of a research and analysis company like ours is to help customers align business objectives with technology. Proper application requirements and system architecture are paramount, as is the correct SLA.
Cloud computing is without a doubt one of the most exciting technologies of the new millennium, plenty of great success stories will result from a decision to use the cloud. Cloud computing enables a radically different kind of economic model, giving large organizations fast ramp up while allowing small companies to compete in a marketplace without huge amounts of infused capital. Any company with a great idea can spend less time to build out and manage infrastructure, leaving more time for product focus and innovation.
There is no question Amazon needs to change the way they communicate with their customers during a crisis, but we shouldn’t let their lack of transparency overshadow what really went wrong.
