Google suffered an outage this week that left 14% of their users headed to sites in China rather than the US. As such, a traffic jam between the US & Asia ensued. Google’s SVP of Operations posted an update to their blog:
http://googleblog.blogspot.com/2009/05/this-is-your-pilot-speaking-now-about.html
The larger question (and one many are asking) is: Should this cause a change to your cloud strategy? The short answer should be no. However, if you’re asking this question, then the immediate question is probably yes. Let me explain…
Many view leveraging of cloud services as a panacea away from the binds of traditional methods. While that is true, there are differences. One difference is how your providers deliver services vs. traditional methods. In your own facility, you would consider redundant options for critical services. Leveraging cloud-based services is no different. However, the redundancy needs to move up the stack. Instead of simply looking at redundancy at the infrastructure layer, redundancy needs to happen at the provider layer. This is a change from the traditional model where you are the provider.
In summary, if your cloud strategy takes into account redundancy across providers, then your strategy does not need to change. An outage at one provider does not significantly impact services to users. However, if services are delivered from a single provider, then yes, your strategy should probably change.