Yesterday, Amazon's cloud services went into free fall as services began to fail and, today, the issue is not yet resolved; Amazon engineers continue to work to fully repair the damage and bring all services back into operation. This incident is a stark reminder that "to the cloud" is really easy to say but organizations can't assume that moving services to the cloud equates with no possibility of downtime. While Amazon's services are among the most mature and solid of all of the vendors out there, even they will have the occasional hiccup, which we've seen over the past couple of days.
The cloud worries me a bit, but mostly because of the hype factor. Not because I have any particular misgivings about the technology. I certainly see the benefit and the allure of simply ejecting internal infrastructure and services, but making such radical shifts must be done with an eyes wide open approach that makes good business sense and organizations need to carefully analyze what services can be appropriately outsourced and which ones have to stay in-house. Further, complete cost/benefit and risk analyses need to be performed for all scenarios – outsourcing & insourcing – and a decision made from there.
Amazon's outage reminds me that not even the biggest players will be perfect. Amazon has certainly done an incredible job over the years in showing people just how solid the cloud can be, but even that solidity can't prevent every possible downtime event.