Data centres, as a rule, have at least one alternative source of power. If power does fail, an uninterruptible power supply – usually battery based – kicks in, covering the gap until a diesel powered generator is online. When the regular power supply is restored, the generator is switched out.
What happened with the BA system is interesting, because the problems appear to have happened when the main power was restored. BA’s chief executive Alex Cruz is reported as blaming the failure on some kind of power surge which damaged servers physically. That’s strange for at least two reasons. One is that any large system would, you might expect, be routinely equipped with power surge protection. The other is that restarting large computer systems is a tricky business; things have to be done in the correct order.
The problem was apparently mirrored at another BA data centre located a short distance away, which raises a further question: how could that happen, particularly when reports say there were no local power problems?
BA is apparently conducting an urgent investigation into why the systems failed. All those responsible for powering data centres and similar large installations will be interested in the findings. Hopefully, BA will publish the results.
Until then, you have to wonder why a company whose global operations are entirely dependent upon computers didn’t appear to have the right support systems in place.
