Why not - same Oracle server data (about 200 GB) compressed by IBM 3583
got approx 2:1 while using node-compression got 3.4-3.5:1.
Another customer case - 600 GB DB (cold backup, no TDP involved) had to go
to DLT and achieved again 2.2:1 compression. Switching to diskpool with
node-compression relieved DLT drives contention and as side effect
improved compression ratio. Compressed data was 130 GB initially and grew
up to 160 GB over months. Thus the ratio was 4.6:1 going down to 3.75:1
but is still much higher than 2:1.
Another usage of node-compression is when you have to backup a lot (Nx
10^5 - 10^6) of small files. Drives cannot stream very well but compressed
data sent over LAN and aggregated by TSM server streams better. Again
performance benefit.
Completely agree with the argument - this highly depends on backup case
and data contents. And I am using node-compression only on mighty servers
- 6, 8 or more processors. The results are very pleasant. You can look
also Dwight's results on E10k - 20 processors can feed even Gb Ether with
compressed data.
Zlatko Krastev
IT Consultant
Mark Stapleton <[EMAIL PROTECTED]>
Sent by: "ADSM: Dist Stor Manager" <[EMAIL PROTECTED]>
13.01.2003 22:32
Please respond to "ADSM: Dist Stor Manager"
To: [EMAIL PROTECTED]
cc:
Subject: Re: Objects compressed by
On Thu, 2003-01-09 at 07:11, Zlatko Krastev/ACIT wrote:
> Yes, usually TSM node-compression gets better compression ratio that
tape
> drive compression.
I'd like to see figures on this. My experience has been that
hardware-based compression is both faster and more efficient than
software-based compression.
--
Mark Stapleton ([EMAIL PROTECTED])
