<img src="http://www.gblwebcen.com/41894.png" style="display:none;">

Volta Data Centres Blog

Power-less? - by Phil Alsop, Editor, DCS Europe Published

Posted by Volta Newsroom on 20-Jun-2017 09:00:00

I sat down at my laptop this morning, clicked the mouse to wake up the machine…nothing. I pressed the start button gently, then rather more firmly – still nothing. Just as I was beginning to do what we all do at such moments – trying to remember when I last backed up the hard drive, I spotted that the power cable had come loose from its socket on the side of the laptop. The cable plugged in and, a couple of seconds later, the start button was pressed, and the laptop was back in business. Time to write the latest Volta blog.


I’ve yet to conduct a full scale enquiry as to just how the power cable became disconnected. It’s stretched just about as fully as it can be to reach from the plug socket to my desk – not a great start. But I’m fairly convinced I’d have noticed any such disconnect yesterday, so I think it’s time to interview our four cats and two dogs and work out which one has the particularly guilty looking expression when I point at the laptop and ask the question. Of course, if one of them does confess, I should also take a fair share of the blame – as with television (never work with very young children or animals), so with computers. I still chuckle when I remember the story a work colleague told me many years ago of how he’d bought his children a brand new Apple Mac computer for Christmas, and then discovered his cat micturating on the said computer on Boxing Day!

Anyhow, I guess all of the above demonstrates the importance of power to the proper running of IT systems and how easy it is to allow human error to cause problems – whether it is our own, or that of other people. Add to this the potential for natural disasters – storms, flooding and the like – to cause power supply disruptions and it’s clear that anyone running a data centre needs to ensure that, as far is humanly, or even robotically, possible, the potential for any power disruption is kept to a minimum. Furthermore, it would be sensible to understand that, no matter what precautions taken, data centre power loss from the grid, with the attendant IT disruption, is going to happen sooner or later.

One has only to google ‘data centre power loss’ or some such phrase to discover a variety of examples of high profile data centres that have suffered from power outages. Whether or not these situations could have been foreseen and avoided, by far the most interesting part of these stories is the consequences suffered by the organisations whose data centres went down. There’s normally huge reputational damage, with an associated financial cost.

So, how could these companies, and your own, avoid finding themselves in such a horrendous situation? As ever, risk assessment is everything. If money is no object, then you can backup/duplicate/replicate/mirror your IT systems and applications all over the planet. Unless an extraordinary series of unfortunate, and unpredictable disasters across the world simultaneously destroy, say, your six data centres, all of which are capable of running your organisation, then chances are that you’ll never be faced with the public relations aftermath of a major IT meltdown.

Conversely, if you have no budget whatsoever to backup/duplicate/replicate/mirror your IT and applications infrastructure, and you know the whole thing is held together by a few elastic bands, and run by your one and only, soon-to-be retired IT expert, then you would do well to heed the horror stories your google search has uncovered, and start taking the issue of data centre power and resilience rather more seriously.

The good news is that, even if you still insist on owning and operating your own IT infrastructure, there are plenty of colos and service providers out there who can provide a very reasonably priced service to backup your applications and both live and archived data, and/or mirror this infrastructure in their own premises, allowing you to failover pretty much seamlessly. And when I say reasonably priced – think of the numbers being thrown around when a high profile organisation suffers a very public, and prolonged IT failure. In the short term, share prices plummet, as do revenues. Longer term, plenty of money has to be spent to try and reverse this trend.

Moving on to the specifics of the power supply itself. Most businesses rely on the standard power supply delivered by a utility provider. While a return to 1970s and the three-day working week and endless power cuts seems extraordinarily unlikely, even in the 21st century, there is the occasional power outage. Whether man-made (yes, utility workers do still dig through cables!) or the result of extreme weather conditions, any organisation which cannot survive without power for more than a few minutes, would do well to investigate the idea of some kind of backup power system. At the same time, it would be a good idea to investigate just how easy it is to re-start the IT systems (however non-essential they might be to your business) after a power cut of only a few minutes.

So, for many, if not all businesses, power resilience is important, if not crucial. Having more than one power feed to a site should get around the problem of someone digging up the road and damaging the power feed, but if all the power is ultimately coming from the same resource – the National Grid – there will be times when no amount of separate power feeds will overcome a power outage.

Time to investigate and invest in backup power – or your own, backup power station that relies on an alternative, maybe renewable, power source. Chances are the budget doesn’t stretch to the second option, so it’s backup power – diesel generators, batteries and the like.

Depending on just how much resilience you want, then the investment in such backup power generation technology will not be inexpensive. And, of course, you will have to regularly maintain and test such infrastructure. The data centre industry is littered with stories of data centres who had backup power, never tested it and then one day they needed the emergency power, and it didn’t work! You might also have to deal with the problem of where to locate this ‘extra’ power source – both in terms of space constraints and, for sure, health and safety issues.

If you think that investing in backup plant that stands idle virtually all year, every year, and is only there to support a relatively small, single IT workload, then it might be worth considering the merits of a colo provider. As previous blogs have, hopefully, illustrated, the colo option can provide access to facilities infrastructure at a fraction of the cost of owning and operating it in-house. Power resilience and backup power offerings are a great example of such value. After all, a colo can ensure two or three separate power feeds into the data centre facility, and invest in the required level of backup power systems, and then spread the cost across the whole customer base – a massive saving for each such customer when compared to having to purchase the necessary power plant themselves.

And I know one or two of you are thinking that some of the reported data centre disasters are due to the fact that a company has chosen to outsource various IT functions – but that just reinforces the point that I have made throughout the blog series: choose your colo and managed services providers carefully. If someone is offering you an unbelievable infrastructure/managed services deal, at an unbelievable price, then it’s probably just that: unbelievable.

Similarly, if someone tells you that they’ve never had any downtime in their data centre, ever, then chances are they are about to experience some (that’s what the stats would say), or the small print describes downtime as ‘an outage lasting more than half an hour’!

Minimising the risk of any data centre power outage is a relatively simple procedure (and my pets are at least under a temporary ban from my home office), but it’s important to understand that you need to plan for such an outage, whatever the cause. Imagine how your customers will react if they cannot access your products/services for a prolonged, critical period of time (and this definition will vary from business to business). Then try and put a price on this downtime. And then work out if you can afford not to take the idea of data centre infrastructure and power supply rather more seriously than you do now. Talk to your in-house power and IT folks, and talk to some colos. Do some more sums, and then calculate your power resiliency options. It just might be that the colo option makes a great deal of financial sense.

Topics: Volta News

Subscribe here!

Recent posts