The first time I heard about Amazon’s EC2 service, I was struck by the thought that very soon a biotech company, with modest computing needs would not need to have a datacenter with a big cluster, but rather use a service like EC2 to satisfy it’s computing needs. A couple of weeks ago I asked Charlie Bell about EC2 and MPI, and guess what, it’s doable. Jeff Barr blogs about EC2, Mpich2 and Blast. Perfect!!! Now, mind you, if you are in the situation where you are crunching the entire human genome all the time, a dedicated cluster is worth it. But if you are a company that usually runs small jobs and occasionally needs the big spike in compute power, the EC2/S3 combo is quite intriguing and I have a feeling it would be fairly cost effective.
One of the advantages for going this route is that a company doesn't have to worry about maintaining a big server farm. As Charlie said when he was talking about it "let us take care of all the muck"
Computing in the cloud
The first time I heard about Amazon’s EC2 service, I was struck by the thought that very soon a biotech company, with modest computing needs would not need to have a datacenter with a big cluster, but rather use a service like EC2 to satisfy it’s computing needs. A couple of weeks ago I asked Charlie Bell about EC2 and MPI, and guess what, it’s doable. Jeff Barr blogs about EC2, Mpich2 and Blast. Perfect!!! Now, mind you, if you are in the situation where you are crunching the entire human genome all the time, a dedicated cluster is worth it. But if you are a company that usually runs small jobs and occasionally needs the big spike in compute power, the EC2/S3 combo is quite intriguing and I have a feeling it would be fairly cost effective.
Further Reading:
Supercomputing 2006
Technorati Tags: Amazon, EC2, Bioinformatics, MPI, BLAST