Yesterday the infamous Twitter “fail whale” reappeared in a very prominent way. Displaced users were forced to Facebook as a result. Adam Martin commented about Twitter’s not so clear explanation in an article on the Atlantic Wire. Twitter went down, it seems, because a data center system failed and so did its backup. Martin didn’t find that explanation very informative, and neither do I. Data centers comprise lots of redundant systems, and Twitter operates, I am sure, multiple redundant data centers. Double fail on Twitter.
Paraskevi Papahristou also failed. Her July 22nd Tweet got the triple jumper removed from the Greek Olympic team before the games even began. Papahristou apologized, but it was too late. All of social media, including Twitter and Facebook, can be considered a platform for change and expressing new ideas, but the platform itself stands in a harsh spotlight.
These issues are extremes of the kinds of “failures” that the internet can bring. There are many other examples. Today’s news included a report that the U.S. critical infrastructure had experienced a 17-fold increase in cyber attacks between 2009 and 2011.
Paraskevi Papahristou also failed. Her July 22nd Tweet got the triple jumper removed from the Greek Olympic team before the games even began.
None of these attacks necessarily involved the physical infrastructure that Mission Critical covers. For the most part, these vulnerabilities seem to exist in the software or user realms. And I tend to think that relative to these other sectors, power reliability in data centers is relatively well managed. In fact, other issues, especially energy, often distract us from the reliability concern. Sometimes we through up our hands, frustrated that IT uses the energy while facilities owns the bill.
And maybe these are two sides of the same coin. IT and end users may own the knotty problems of system failure and downtime, but should facilities be glad that it is so? Well, we know how we feel when that unexpected rack arrives at the loading dock.
Siloed approaches to problem solving tend to fail in business, and I think it’s also true in the data center. Will anyone at Twitter feel real secure if the fail whale starts making regular appearances again? I think not, no matter what brings the whale back.