I was just reading Managing risk in the wake of Amazon's cloud outage and it discussed the outage that occurred last week where the cloud servers at Amazon went down. The issues were said to occur due to bugs in the code or inclement weather, depending on which source you read. With that being stated, here are my thoughts on the dependability of cloud computing. In the early 2000s I was working in corporate america for a variety of companies because I was a web developer consultant. I had the opportunity to see a variety of organizations IT structure and how they reacted to the monthly and even weekly server issues that occurred within the company. Microsoft Outlook email would go down, Intranet server would go down, Internet connectivity would go down and every time a vital service went down, no one panicked or questioned the use of the service, they just waited until it was resolved. Yes, millions of dollars was being lost in man hours due to the lack of productivity during this downtime but there was never a mention of removing the Microsoft office servers or any other servers as our solutions.
Now, since the cloud is a viable solution to help small businesses compete with larger corporations and grow from a one man website into a fortune 500 company, it appears to come under attack more often than not. The cloud, and more specifically Amazon cloud solutions and even Google’s cloud solutions have a very successful uptime rate that is upwards of 95% - 98%, that’s extremely higher than what we dealt with in the early 2000s and for a fraction of the costs. When Netflix, Instagram and Pinterest went down last week, it wasn’t the cloud that we were upset with, it was the fact that these services that we depended on went down and someone had to take the blame. When in reality, we could’ve all been patient and waited a few hours or a day for Amazon to fix the issue and we go back to life as normal.
Companies need to make sure they have redundancy plans in place and backup sites available but the best lesson for working in the cloud was the title of this article - How to deal with cloud failure: Live, learn, fix, repeat.