With redo transport compression, we may stop considering compression at hardware or Cisco switch level, which only a few companies in the world are doing. But it would be nice to check the CPU overhead of this compression. If it’s too much, compression with dedicated hardware may still be desired.
This is the weblog for Eagle Fan, a principle DBA working for ebay.com.
You can contact me at eagle.f@gmail.com with feedback and comments on this site.
With redo transport compression, we may stop considering compression at hardware or Cisco switch level, which only a few companies in the world are doing. But it would be nice to check the CPU overhead of this compression. If it’s too much, compression with dedicated hardware may still be desired.
The cpu overhead depends on redo generation rate and what the data is in redo. I didn’t see obvious overhead while doing test.
Have you already upgraded your production database to 11g?