While the pundits debate the minutiae of cloud computing, and spend hours in detailed and complex arguments of what should and shouldn’t be moved to the cloud, other people are simply going out and doing stuff on the cloud that would previously have been almost impossible. It’s always nice to come across these case studies because, if nothing else, they allow people to create an arsenal of “look, proof that cloud is actually a thing” stories. I came across one of these examples the other day on the CycleComputing site. Yes it’s a vendor site and yes it’s self service, but look beyond the vendor specific story for awhile and thing about what the users actually achieved here.
It’s worth reading the post but for those without the time, here’s the synopsis:
CycleComputing were contacted by Opscode to scale-test chef. One of CycleComputing’s science clients was running a massive scale run against a cancer problem – something to do with simulating the effects of different compounds on a protein associated with cancer. The run was estimated to take 341700 hours. CycleComputing built a utility supercomputer with some 10600 cloud instances, each of which were an individual multi-core machine – apparently this is the largest cloud HPC environment ever built. If it had been built physically it would have required 12000sq feet of data center space and cost $44M. Instead, over a two hour build time, and a nine hour run time, the total cost of the job run was $4362. 39 compute years, spun up in only a couple of hours, and completely run in half a day. Compelling story huh?
Of course it’s easy to say that this sort of requirement is very much an outlier and that’s a valid argument. But it seems to me that as we move to a world where we’ll be facing far larger computational jobs on a ever more frequent basis, the opportunities for cloud to enable something that has never been possible, or at least viable, before is also increasing.
Cloud’s not for everything, and there’s any number of workloads that will remain for the foreseeable future where they are. But for the sort of workload that CycleComputing described above, cloud is truly invaluable.
Great example of the cloud in action.
Great example of scaling, yes this is an outlier, but I suspect as systems evolve even in such mundane things as ERP that such will become more the norm. I can remember a few years back when it was reasonable to think a 10GB database was big for a SME business system database. Now it seems more like the norm or even small