Excellent article on the work of the Bioteam at Sun Life Sciences.
Some really choice quotes here, including answers to the perennial "why do we need local resources when we have e.g. the NCBI" type questions, often to be heard around our school...
Some of those quotes in full:
"A lot of people in the (life sciences) industry have purchasing authority for perhaps 25,000 or 30,000," [Dagdigian] says. "Thanks to the power available from desktops or small server systems, for that sort of investment you can put together a six- to eight-node cluster, plus a storage array. With some creative packaging, that becomes a personal 'sandbox' that you can deploy in someone's closet, not in the data center."
As an example of how these affordable, production-ready solutions can benefit researchers, Gloss cites the process of running BLAST (basic local alignment search tool) searches, or other types of genomics queries through the online database at the National Center for Biological Information (NCBI). "The site provides limited throughput, and the turnaround time for getting results is very slow," he says. "Now that firms are installing a local cluster with few nodes, and instead of waiting minutes or hours for result sets from NCBI, they're generating their own results in minutes or seconds."
"An organization's research staff can analyze data according to their own schedule, making them more productive and allowing researchers to pursue their objectives more effectively. At the same time, the organization's IT staff doesn't have to support users on an individual basis
