Bladewatch.com

Chris called me up to tell me what he was working on and ask me what I was up to. Chris is currently continuing to save servers in his role for one of the investment banks in Canary Wharf and he has been learning DataSynapse (one of the technologies which I remain a fan of). Anyway he had to put together a wiki page talking about the benefits of Grid as the IT team were trying again to explore Grid computing and more effectively see how it could be utilized to reduce IT costs and improve utilization of the platform. With that in mind as he asked me what I was up to (unracking servers, you’d be surprised what you can learn about server, kvm and cabinet design doing that) and I thought in response I’d blog my comments, which as ever were of the top of my head with no pre-planning.

Share compute resources – achieve more with less, have your batch run on systems which might otherwise be idle when you need them

Avoid further capital expenditure – improve the utilization of your server estate, have 300 servers run at 90% utilization rather than 800 running at 24%

Consolidate workloads around the application based on capacity and geographical requirements – you might have Singapore run their end of day just before London opens for business

Improve reliability – this is more around the shared and distributed engine/workload space – with the right infrastructure and application code in place, there should be no reason why server 19 failing should result in disruption to service, to your SLA or your end user

Achieve what might not be possible without shared processing – that batch which takes 24 hours to run could (with the everything in place) run in four hours if we could distribute workload more effectively.

No related posts.

Related posts brought to you by Yet Another Related Posts Plugin.