The below image shows what happens in essence when you are using the Oracle database Smart Flash Cache options.
- When a block is retrieved from the storage it is stored within the buffer cache of the system global area.
- When using Smart Flash Cache a block is not removed from the buffer cache however is evicted to the flash cache instead.
- When a block is needed again and is not available in the buffer cache in the SGA it is retrieved (if available) from the flash cache instead.
By implementing this you can avoid, up until a certain level, recurring calls to your storage device. Requesting a block from storage is slow and if this can be avoided the benefits to the performance of your database are directly visible. Next to this, and often overlooked, is the fact that this is also resulting a positive effect for other databases and applications that make use of the same shared storage device. Due to the fact that you can lower the I/O operations on your shared storage there is more room to handle the requests from other applications and due to this there can be a positive performance gain due to this indirect relation between the components.
One good thing to keep in mind is that this option is only working when your database is deployed on Oracle Linux or on Oracle Solaris.
One good thing to keep in mind is that this option is only working when your database is deployed on Oracle Linux or on Oracle Solaris.
For more information please do have a look at the presentation on this subject which is embedded below:
No comments:
Post a Comment