I’ve had lots of conversations over the years about the effectiveness of Uninterruptible Power Supply (UPS) arrangements. Opinion has been divided and for sure there are issues and different theories to consider. Where do you stand on this?

I’d love to hear any stories illustrating where UPS has saved the day – feel free to let me know – but it seems that in many situations the existence of a UPS capability has added only limited or no value so is it really worth the money?

UPS typically forms part of an overall solution. When it is designed, implemented and maintained correctly, for instance as part of a back-up generator capability it can be effective and therefore worth the investment. The key is to ensure the UPS is linked to the right kit (i.e. not the TV in the staff canteen for example!) and can keep it running long enough for the generated power to reach the required operating level and take over. Test it guys!!!

The devil is in the detail but this kind of solution makes sense to me – I can see how this would work.

But what if you don’t have a back-up generator? What do you expect the UPS to do for you?

Faced with this question the response from IT Directors is typically that the UPS allows sufficient time for the kit to be shut down ‘gracefully’ or in a ‘controlled manner.’ This is fine if your team can execute a shutdown within the time available i.e. the duration of the UPS power. In the worst case scenario, for instance where the engineers need to travel to site to perform the shutdown you may run into some problems. The clock starts ticking when the mains power is lost remember.

…..and actually does it really make any difference?

If my laptop or phone ‘dies’ through lack of power, the devices seem to come back up OK when a power supply is restored, irrespective of the state they were in when they stopped. It’s almost as if they perform a series of health checks and repairs on the way back up. Surely servers and other IT kit does the same? Doesn’t it?

I worked with a datacentre manager at one large organisation who firmly believed in this theory and was all for simply ‘pulling the plug’ on the datacentre to test what would happen – his bosses didn’t share the same risk appetite as him though, so unfortunately he never got the opportunity to test his theory. Have any of you been brave enough to try?

A UPS capability may also save the inconvenience arising from short term power outages so I can also see some benefit in such a scenario. We just have to hope the electricity supplier fixes the problem before the UPS expires.

So the conclusions seem to be that UPS does have a place as part of a wider power strategy but may otherwise only be effective as a standalone solution for very short term outages.

Whatever your circumstances and requirements the use and application of a UPS solution needs to be carefully thought through and well implemented. Once it’s done be sure to test it regularly. If you have a generated power solution, this must be tested on an ‘end to end’ basis to ensure all the components work.

January 27, 2016 at 11:24 am
Return to All News
Category: Resilience
Tags: , ,