As I understand it, Isilion is built up from "bricks" that have storage and compute power. They replicate files amongst themselves, so that for every IO request there are multiple systems that could respond. They are interconnected by an ultra fast fibre backbone. So, depending on your topology, it's possible to get a lot more throughput by working on different sections of the same file from different physical computers. I haven't delved into BGZF, so I can't comment. My approach to block GZIP was just to concatenate multiple GZIP files and keep a record of the offsets and sequences contained in each. The advantage is compatibility, in that it decompresses just like it was one big chunk, yet you can compose subsets of the data without decompressing/recompressing and (as long as we actually have to write out the file subsets) can reap the reduced IO benefits of smaller writes. John Duddy Sr. Staff Software Engineer Illumina, Inc. 9885 Towne Centre Drive San Diego, CA 92121 Tel: 858-736-3584 E-mail: jduddy@illumina.com -----Original Message----- From: Peter Cock [mailto:p.j.a.cock@googlemail.com] Sent: Thursday, October 06, 2011 8:16 AM To: Duddy, John Cc: Greg Von Kuster; galaxy-dev@lists.bx.psu.edu; Nate Coraor Subject: Re: [galaxy-dev] Tool shed and datatypes On Thu, Oct 6, 2011 at 3:48 PM, Duddy, John <jduddy@illumina.com> wrote:
I'd be up for that something like that, although I have other tasking in the short term after I finish my parallelism work. I'd rather not have Galaxy do the compression/decompression, because that will not effectively utilize the distributed nature of many filesystems, such as Isilon, that our customers use.
Is that like a compressed filesystem, where there is probably less benefit to storing the data gzipped?
My parallelism work (second phase almost done) handles that by using a block-gzipped format and index files that allow the compute nodes to seek to the blocks they need and extract from there.
How similar is your block-gzipped approach to BGZF used in BAM?
Another thing that should probably go along with this is an enhancement to metadata such that it can be fed in from the outside. We upload files by linking to file paths, and at that point, we know everything about the files (index information). So need to decompress a 500GB file and read the whole thing just to count the lines - all you have to do is ask ;-}
I can see how that might be useful. Peter