First up today is reading about ARM morphing it’s mobile phone CPUs into low power data center servers. I think it’s already way too late for this, much less in the 12 months they’re predicting to ramp this up. I just don’t see how underpowered low energy using dedicated servers are going to match up to the efficiency gains of virtualization and cloud computing on servers with decent CPU (I have the same thought about Intel’s Atom servers too).
Next is the piece about finding and resolving the weaknesses of data centers. Single points of failure aren’t necessarily the problem, not correctly figuring the risk of an SPOF is. Even so, I think the biggest risk in the data center is still lousy operating procedures (such as doing unnecessary manual “maintenance” inside of live electrical equipment). It doesn’t matter how redundant the data center is, it can’t protect against people doing dumb things.