Last week's outage for some of Amazon's cloud computing customers has highlighted the importance of planning for cloud failures. While the large data centers run by the most competent companies such as Amazon, Google, and Microsoft have enviable performance and reliability, they are not perfect. Customers of cloud computing must still reckon with backups, failover, and contingency plans if they want to ensure a high level of availability.
I predict that we will witness a significant number of cloud computing failures which will affect large and small companies, governments, and individuals. Some will be from fumbles by the major players, but an even larger number will come from new companies and dabblers in cloud computing. The whitehot hype around cloud computing will create large numbers of businesses that ultimately fail.
Cloud computing is creating a smaller echo of the dot com boom, with small companies armed with ambitious business plans trying to carve out their pieces of the cloud pie. The frontier mentality will lead to aggressive marketing, aggressive sales, and aggressive pricing.
What will happen when new cloud companies start to run out of capital, or their venture capital masters give them a tighter leash? Will their priority be to help customers find a safe landing? How many cloud providers are willing to give customers backups that could quickly be restored with another provider?
The implication for customers of cloud computing is that customers must maintain options for moving from one provider to another or even moving systems in-house if necessary to protect themselves from technical and business contingencies.