This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Local Clovr Cluster Or Ec2?

I'm setting up a annotation pipeline for some prokaryotic Illumina data, and I thought CloVR seemed like a good place to start. I know it's still in beta, but hopefully someone has some experience running it. I have two questions:

  1. Is it possible for CloVR to operate on a small group (3-5) of local machines?
  2. If not, which EC2 instances offer a reasonable cost/time ratio? For instance, do the Micro instances have sufficient power to serve as nodes in the BLAST/HMMer steps? From a financial standpoint, making the most out of the free usage tier would compensate for up to twice as much additional processing time, relative to other EC2 instances.

Thanks.

edit:

I noticed that the CloVR AMIs are instance store only, but I think that the conversion to EBS should be trivial.

illumina next-gen sequencing cloud
• 2,588 views • link
updated by Mndoci • Mndoci • written by James Estevez • James Estevez

1 answer

You shouldn't use Micro for any workload that requires sustained throughput. They are designed to serve a few requests a minute and disk IO is going to be limited.

Converting to EBS should be easy, and if you do so, consider using HVM and using the CC1 instances. If you don't need memory,C1 instances have the best price/perf. If you do need memory and some IO, tough to beat CC1.

• 1 views • link
written by Mndoci • Mndoci

Log in to answer this question.

More posts like this