Steve Hone CEO, DCA Trade Association

Service Availability and Resilience

This month’s theme is Service Availability and Resilience. It’s only natural that every data centre wants to ensure they are as resilient as possible. Data centre owners spend hundreds of thousands on technology trying to achieve this goal and then tens of thousands to keep it maintained both from a support and power perspective. The other holy-grail that everyone seems to chase is the lowest PUE figure possible.

These two objectives are actually diametrically opposed to one another as the more resilience you build into your data centre the more inefficient it becomes as you need more energy to maintain the infrastructure you have running. Thus often has the perverse effect of sending your PUE up, not down. It is worth noting at this point that the “E” in PUE stands for “Effectiveness” NOT “Efficiency” which is a mistake often made and this yet again puts a different prospective on it.

On the subject of effectiveness, it is also worth noting that just because you think the facility you run or use is resilient on paper (e.g. N+ this and N+ that) please don’t automatically assume these measures will actually be “effective” in ensuring your service remains ‘Available’. Making sure you have fully tested processes and that a business continuity plan is in place is just as important, and some would argue more important, than the hardware itself.  After all, when things go wrong, and they will, it is normally human error which is ultimately to blame.

One very real example of this happened recently to one of the largest cloud providers to the insurance world, having suffered a major service outage at its third party colocation data centre despite boasting a 99.99999% uptime record, all eyes quickly turned to the data centre provider as the guilty party. I am sure there will be lessons learnt by the colocation provider, however I can’t help wondering if the ultimate reasonability for this outage actually lies with the managed service cloud provider for not asking the right questions and not planning more effectively for the worst.

Having a regularly tested business continuity strategy with your suppliers is critical, in fact it is often referred to as an insurance policy to help protect your business and your clients. The irony of this story is that it involved one of the largest providers of cloud based services to some of the top global insurance companies – who all failed to recognise the risk or value in investing in an insurance policy of their own.

It would be unfair to single out this one incident, provider or sector as I’m sorry to say this story is not uncommon, there are clear lessons to be learnt here for all clients seeking cloud based services and for all cloud providers seeking a hosting provider to deliver and underpin their offering.

Problems do occur irrespective of how resilient the provider says their service is and when it happens the SLA won’t save you.  It is vital you do your due diligence, ignorance is no defence in the eyes of the law and no defence in the eyes of very unforgiving clients.

I would like to thank all those that contributed articles this month. Next edition is an opportunity for your customers to speak for you in the form of client case studies so if you would like to submit then please forward them to Kieranh@datacentrealliance.org

Deadline for submissions is the 15^th March.

Finally, The DCA has an update seminar the afternoon before DCW 2017 at the Excel on the 14^th March, if you are a DCA member or someone interested in finding out more about the Trade Association and the value it can deliver, then you are more then you are more than welcome to register and attend. Full details are available on the DCA website www.datacentrealliance.org

.