[Segway-users] Segway on Amazon AWS

[Jay Hesselberth]

Dear Segway users,

Michael asked me to put together some instructions for getting Segway
(and genomedata & segtools) running on Amazon Web Services.  This
turns out to be straightforward to do, and I've outlined the process
here (http://sc3id.ucdenver.edu/jhessel/segway-cloud/segway-amazon-cloud.html).

Essentially it amounts to:

1.  Signing up for an AWS account (you'll need a CC number)
2.  Installing starcluster on your local client
(http://web.mit.edu/stardev/cluster/)
3.  Copying the public segway snapshot over to your own EBS volume
4.  Launching a AWS cluster instance of arbitrary size with
starcluster, with specific instructions via starcluster to mount the
segway volume (instructions above)
5.  Transfer some data in (e.g. a genomedata archive) onto your
cluster (e.g. via scp).  You'll want to create a separate EBS volume
for this so that the data is persistent (I typically make 50 Gb
volumes as I need them).
6.  Launch a segmentation.

We have had good luck running multi-day segmentations for reasonably
cheap.  By the way, I was mainly motivated to do this because the
cluster that I have routine access to is a Mac OS X cluster, and it
turns out to be non-trivial to install segway in that environment
(mainly SGE/drmaa limitations).  I've also figured out a workaround
for that, if anyone is interested.

Best,
--
Jay Hesselberth, PhD
Biochemistry & Molecular Genetics
University of Colorado School of Medicine
off: (303) 724-5384
lab: (303) 724-5486
jay.hesselberth at gmail.com