Andrew> while i am no stranger to large data, i have found myself
Andrew> outside my comfort zone at work. the department i have joined
Andrew> (i lateralled within research) has traditionally used Sun as
Andrew> mid-range storage and Hitachi as their high-end.

Andrew> we now need to look at multi-PB disk systems (say 2-5PB),
Andrew> which i have always thought of as a different market, with
Andrew> players like Parnasas.  as normal, we care rather less about
Andrew> IOPS and more about bandwidth and $ per TB.

What are your requirements for this storage?  Do you want just one
large filesystem spanning it all?  Do you need multiple clients
accessing the storage in a streaming mode (as I suspect from your
comments) or are you looking at a more random access workload?

I've been a Netapp user for a long time, but they won't scale anywhere
near this level at all.  I've been talking with Isilon and they really
tempt me alot.  I especially like how you can just add storage by
dropping chassis into their Infiniband cluster backend and it just
grows without you having to do anything.  Of course once you hit their
switch limit, it's a bit harder to expand I suspect.

Curious about your backup strategy, and reliability needs too.  

I know this is archaic technology, but would some sort of tape library
backend fronted by and HSM solution work for your data and workflow?  

Any chance you write another Usenix/LISA paper describing your work?
Always fascinating to read about large deployments like this.

Cheers,
John
_______________________________________________
Tech mailing list
[email protected]
https://lists.lopsa.org/cgi-bin/mailman/listinfo/tech
This list provided by the League of Professional System Administrators
 http://lopsa.org/

Reply via email to