Transcript

Smart Scan in Theory

The first feature I want to look at is Smart Scan. The Smart Scan capability allows the Exadata layer to take over a vast amount of the processing, the workflow to processing SQL, the offload, the workload, from the compute nodes to the cell nodes.

What can be offloaded? Row selection and column projection, conventional storage will do block serving. Your session issues a request, the storage system delivers the blocks in database buffer cache all of the dynamic regions of the PGA, and your session then has to filter the rows, extract the columns.

With the Smart Scan, the storage tier does not do block serving. The storage tier itself does the row filtering and the column projection and returns only the relevant columns, the relevant rows directly into the PGA of the session that would run the query. We bypass the buffer cache entirely. In addition to that and as part of the Smart Scan, many, many functions can also be evaluated by the cells, so I’ll show you a list of which functions can be evaluated shortly.

Of particular importance for the data warehouse application, simple drawings can be implemented within the cells, not all types of drawings, but simple drawings of fact tables, two-dimension tables. They can also be implemented in the storage tier.

The result of this is a vastly reduced amount of data traffic into the compute nodes. We do not have to flood the database buffer cache with many, many millions of blocks when perhaps all we need is a single scale or value. The CPU usage in the compute nodes is greatly reduced, particularly through offloading function evaluation, and the end result for the user is a terrific enhancement to response times.

×
Free Online Registration Required

The tutorial session you want to view requires your registering with us.

It’s fast and easy, and totally FREE.

And best of all, once you are registered, you’ll also have access to all the other 100’s of FREE Video Tutorials we offer!

 

×