09-13-2007 03:20 PM
Let me start by saying that after spending years looking at many data dedup products (of which none integrate well with the backup applications), I have high hopes for this Puredisk product.
From all that I have been able to find, this code runs on a Linux box, that would be SAN connected to the storage system (assume a large 100TB high performance box). In that we have a lot of LTO2 and 3 drives running now (and are certainly going to go to LTO4) , my concern is that there may not be enough bandwidth in the box, even if it supports multiple FC HBA's between the SAN/Media servers and the SAN storage system. So what is the performance and scaleability? Do you replace a single LTO3 drive with a Puredisk system for each one? How does it scale in real life?
I am trying to understand the sizing of this box (and hardware/licensing costs) so that I can plan the future upgrades to Gen4 drives, as they simply can't be fed from anything else but disk, of course.
01-10-2008 02:12 PM
02-04-2008 08:30 PM
04-14-2008 08:01 AM
09-16-2008 12:23 AM
NOTE a few things that have changed since the original post. In Puredisk 6.5:
A Puredisk Content Router (CR) can handle up to 8 TB of data.
A Puredisk MetaBase Engine (MBE) can handle up to 100 million file versions.
A Puredisk Storage Pool can have multiple CRs and MBEs, so to really answer the first post, Puredisk can handle, and load balance, across multiple servers, if that was a concern. Additionally, you can always add PD servers into the mix later to get over these number restrictions, and performance restrictions. Adding additional CRs will automatically have load-balancing happen across them. Adding MBEs would allow for load balancing of the processing of meta-data.