Shrink wrapped grid computing
As we get more and more used to parallel processing and multi-core architectures I’m noticing an encroaching tide of technology threads that are encouraging us to consider High Performance Computing and Grid Computing as more easily accessible options for intensive computing environments.
High Performance Computing and High Performance Technical Computing in my experience are cluster-based highly complex propositions not undertaken lightly. They typically suit the financial services and defence sectors as well as laboratory-driven research projects. But, as is the way with the technology industry in general, the option for this power to be made more accessible for any type of solution is something that more than one cluster and grid management vendor is keen to convince you of.
Are data intensive applications across a grid becoming more accessible? Are we really seeing a reduced level of complexity for enabling enterprise-class resource sharing and availability? Will companies soon be able to build and deploy models for the grid with greater ease? If they are deployed, will these apps run with the claimed speeds (usually something like ten times faster) they are purported to boast? Will developers be increasing encouraged to develop skills for this space? I’m not sure – but there are plenty of companies out there trying to push this idea. Is it dropping in your inbox yet? It’s in mine, that’s for sure.
NB – SUB NOTE: If you’re wondering where I’m getting all this grid malarkey from, I think it’s interesting to note that it’s not just the usual suspects. If I said, “tell me everything you know about grid computing in 10 seconds…” you’d almost certainly mention IBM, maybe Oracle and Sun, you’d probably say “clusters” and if you’re a fan of bbb.co.uk/news like I am you might mention the grid computing project launched by the BBC, the Met Office and others back in 2003 which aimed to get 50,000 users to volunteer their home computing power in the name of climate change.
Today though, it’s independent companies such as Platform Computing, community-driven bodies such as the Open Grid Forum, web pioneers such as ebay and projects supported by the European Commission's IST programme such as XtreemOS (which aims to build and promote Linux-based support for next generation grids) that are coming to the fore. Is grid computing about to enjoy a renaissance then? Definitely maybe I’d say.