Re: Exadata Cacheflash Compression
- From: George Leonard - Business Connexion <George.Leonard@xxxxxxxxx>
- To: "dmarc-noreply@xxxxxxxxxxxxx" <dmarc-noreply@xxxxxxxxxxxxx>
- Date: Thu, 8 May 2014 12:40:32 +0000
Hi Thomas
As I always look at it, what is faster... is the right answer... => reading X
blocks of data stored in less blocks (aka the physical IO time, be it from
round brown spinny of nice faster flash cache) or the additional CPU cycles to
compress/uncompress that block, the extra CPU cycles to work through the code.
Yours Sincerely
________________________________________
George Leonard
Oracle Engineered System Specialist
Mobile: +27.82 655 2466
eMail: george.leonard@xxxxxxxxx<mailto:george.leonard@xxxxxxxxxx>
Web: http://www.bcx.com<mailto:george.leonard@xxxxxxxxxx>
[cid:6EA60917-E50D-422D-BF1F-E8C59EADC861]
On 08 May 2014, at 1:29 PM, Thomas P S
<dmarc-noreply@xxxxxxxxxxxxx<mailto:dmarc-noreply@xxxxxxxxxxxxx>> wrote:
Dear List,
We have X4-2 Half Rack with 20TB flash cache with database size 8TB. I am very
reluctant to enable flashcache compression as the size of the database is half
of flashcache, but Oracle says enabling flashcache compression is a best
practice. Every code execution need time, even though FC is implemented in
hardware level compression, still I believe, there will be a penalty. In this
scenario, enabling flashcache compression is a wise idea? Any one got
falshcache I/O performance test results with compression and with out?
Note: License is not an issue as we have Advanced Compression license.
Thanks in advance,
Thomas Saviour
Other related posts:
