Don’t buy hardware
We all love it. Hardware has flashing lights, cool lines, and looks really good in racks, humming away. It impresses strangers; validates us as experts. It’s also wildly expensive and fantastically underused. Google (yes, that Google) discovered on the early 2000’s that their data centres were running at 10% of their potential maximum. Much of the UK public sector is a lot worse with the Cabinet Office suggesting that some of us get as little as 5% utilisation of our most expensive assets.
Ask any IT Head why they spend gazillions on hardware and they’ll say “well, it won’t work without it”, or if the IT Head has an MBA he’ll say “the hardware is central to our mission”. Which raises an interesting question – why do we feel it necessary to buy the means of production for computing when we have no qualms at all about buying-in electricity, or water for that matter?
Back in September over 60 IT specialists, and me, met at WAO seminars in both North and South Wales to share good practice (viz, ‘nick ideas’) and, hopefully, have their preconceptions challenged. I like a good challenge to my preconceptions and the scepticism around Cloud Computing was not so much challenging my preconceptions as mugging them.
There were a lot of reasons why not to adopt Cloud. There’s security. Everyone will have access to your data. There’s regulation, the Americans will arrest your Chief Executive……. and, well, there’s no flashing lights are there? None of this is necessarily true. Of course you can mess up security on the cloud in the same way you can leave your own security in a hash if you try, but there’s no real reason why you should, particularly.
I like easy computing. I like to provide a service that just works. I don’t want to run a datacentre any more than I want to run a power station or manage a reservoir. For me, Cloud Computing offers a chance to buy just as much as I want, when I want it. Period, as the Americans say. No more capacity than I need.
This isn’t easy for us as a profession. We tend to think of IT as being an exercise in hardware management, rather than service to customers.
Here’s a thought exercise. Imagine your Chief Executive has banned the purchase of anything physical at all. Server provision is easy to purchase through the cloud (and really easy for the public sector through g-cloud), connectivity and firewalls – tick, software as a service – tick, end user devices? More problematic but a good BYOD strategy should see to that. In fact, if you’ve got decent mobile signal coverage (and you can even get around that) you don’t need the building network either.
Of course, little in life is simple, and that isn’t exactly simple either. But it’s worth considering, how much will it cost me if I don’t buy this upgrade but buy it as an on-line service?
It costs a lot to keep all of those servers warm, fed, stroked, and loved. It might sound like heresy but could you do better, cheaper, with fewer commitments if you bought the service through the cloud?
Not that I’m mildly obsessed, or anything, but on 13 December I’ll be discussing cloud computing: the myths, busted, with Peter Middleton from the Cabinet Office’s g-cloud programme. 2013 has been a pivotal year for public service, a pivotal year in recognising that we cannot continue to deliver under the current model. We need to be smarter, more collaborative, cheaper, simply better at what we do. Time to ditch some myths.
– Evan Jones, Welsh Government