As I sit here watching the snow fall at 6,500 feet in beautiful Lake Tahoe, the peacefulness has allowed me to formulate some conclusions on the exciting events of the past week. Everyone is talking about Amazon’s EC2 outage...
“I said the [network] is the Computer, I did not say it had 100% uptime.” - Scott McNealy (via Twitter)
The story has been everywhere, including CNN; you know when a tech story hits the major news sites it’s a big deal.
Amazon is a phenomenal company, and it is innovating at a pace that even the biggest technology companies cannot compete against. The real problem is that a lot of customers fail to realize that there is one truth to always remember concerning any technology, especially cloud computing - at some point, it will go down. We must help educate our customers so they understand all risks.
The first question everyone asked this week was, “What happened?” quickly followed by, “Why did this happen?” It really doesn’t matter what or why. The only thing that matters is that it was down. I’m not going to play Monday morning quarterback as to why Amazon failed or try to decompile its systems’ architectural choices. There are going to be a plethora of case studies written in the coming weeks about how all of this could have been avoided. The front page story is that the biggest cloud computing company in the world took a massive outage and its customers have put a lot of faith into it just working.
There are a great deal of design decisions that go into an architecture to make it “just work.” Resiliency. It all starts with making the environment as highly available as possible. As an Architect here at Carpathia, I come up with complex solutions daily to make sure that my customers are protected from outages. We work with our customers to design solutions that work for them.
One of my good friends works for a custom home builder. She works relentlessly, making sure her customers understand how and where every wire and every 2x4 will fit into their home. At Carpathia it is no different. This week we had a very engaging meeting here in California with a potential customer to discuss architectural options regarding open source storage products such as Gluster and MogileFS, versus the normal big name storage vendor’s hardware. We are extremely passionate about helping our customers build an environment that will be as redundant as possible and making sure they understand every minute detail. The home that we build for our customers' businesses will be custom, sometimes complex, but always reliable.