[Bioclusters] Designing a life sciences computing facility

chris dagdigian bioclusters@bioinformatics.org
Thu, 12 Dec 2002 18:39:18 -0500


I started to reply but Joe hit it on the head exactly :) The key is to 
overbuild the physical facility, network core and storage/backup 
infrastructure. The compute systems can be aquired and scaled far more 
easily than the foundation stuff.

-Chris

Joseph Landman wrote:
> I just ran through a very similar particular project.  There are many
> things to consider, but let me try to make a specific recommendations:
> 
> 1) build the physical infrastructure to allow for growth
> 	. build enough electrical and cooling
> 	. build enough floor space
> 	. use correct fire suppression
> 	. use thermal kill switches
> 	. ups and power conditioning
> 	...
> 
> 2) build the network infrastructure to scale
> 	. use the right backbone(s)
> 	. get good network connectivity to the outside net (highest speed to
> edge of net)
> 	. redundant external connections
> 	. redundant paths internal
> 	. redundant switches/fabrics
> 	...
> 
> 3) build the computing infrastructure to be scalable and secure
> 	. use the right network infrastructures
> 	. plenty of fast storage and appropriate backup
> 	. use appropriate technology where needed,
> 
> I could go into more depth, but usually you want to look into hiring an
> organization to do this (full design).  Contact me offline if you need
> names of companies that do this.
> 
> Joe
> 
> On Thu, 2002-12-12 at 18:10, Joel Dudley wrote:
> 
>>        How would you design a computing facility for life sciences if
>>you got the chance to build one from scratch? That is the question I
>>am going to be faced with shortly. As our new research building is
>>being built I have a chance to influence the design of the computing
>>facility. So, I would like to ask the members of the list for for
>>their opinion on this matter. The facility will play host to various
>>applications such as image analysis, microarray analysis, sequence
>>alignment, phylogenetic modeling, etc. All suggestions are
>>appreciated. Thanks.
>>
>>Joel Dudley
>>Faculty Research Associate
>>Arizona State University
>>Center for Evolutionary and Functional Genomics
>>http://lsweb.la.asu.edu/skumar/

-- 
Chris Dagdigian, <dag@sonsorol.org>
Independent life science IT & informatics consulting
Office: 617-666-6454, Mobile: 617-877-5498, Fax: 425-699-0193
PGP KeyID: 83D4310E Yahoo IM: craffi Web: http://bioteam.net