As 4,200 disgruntled holiday goers, trapped on the ironically named cruise ship Triumph, finally end their 5 day ordeal, it serves as a reminder that the eggs can have more stake in the state of the basket than the basket holder does.
From the point of view of the cruise line, each booked up ship represents a concentration risk, containing thousands of human beings, their fate, indeed their very lives, dependent upon the correct functioning of a very large and complex system. From the point of view of the passengers, a cruise ship represents recovery risk.
While cloud computing has been relatively smooth sailing for the majority of its passengers, there have been multiple multi-day incidents that required a recovery process of uncertain duration, with ambiguous hopes for success. There have even been clouds that ran aground in the shallow waters of a highly competitive marketplace, leaving their passengers permanently stranded.
Most cloud service providers are able to weather a single packet storm, returning to operational status and compensating their customers with credit for time lost, and maybe even a bit of extra credit. For those who haven’t had their enthusiasm permanently squelched by 5 days without toilets or dinner, the cruise line is offering free cruises to the victims of this mishap. Many of the unfortunate Triumph passengers have lost a week’s worth of vacation—something that they can never recover. Likewise, when a cloud fails, thousands of customers are likely to experiences forms of loss that cannot be compensated for.
Both the cloud and cruising industries have proven relatively reliable, but failures do happen. One lesson that cloud customers can take from a series of vacation-ending fires and floodings is that when a single incident simultaneously impacts thousands of customers, the recovery will be slow and frustrating, and the provider will have no way of compensating their customers for their lost time.