n
Table and Index scans: Scans are performed
inside Exadata Storage rather than transporting all data to database server.
n
Predicate Filtering: Only requesting rows are
returned to the database server rather than all rows in a table.
n
Column Filtering: Only requested columns are
returned to the database server rather than all table columns.
n
Join Filtering: Join processing using Bloom
Filters are offloaded to exadata server.
n
Smart scan can also be applied to encrypted and
compressed data.
Smart Scan
Requirements:
Smart Scan optimization is a run time decision. Smart scan
is not governed by optimizer, but it is influenced by results of query
optimization.
n
Query specific requirements
---- Smart Scan is only possible for Full
Table or Index scans
---- Can be used only for direct path
reads. (A
direct read is a physical I/O from a data file that bypasses the buffer cache
and reads the data block directly into process-private memory.)
n
Additional general requirements
--- Smart Scan must be enabled with in
database. CELL_OFFLOAD_PROCESSING parameter controls smart scan. Default value
is TRUE. Can be set dynamically using alter session or alter system.
---- Each segment being scanned must be on
a disk group that is completely stored on exadata cells.
Situations Preventing Smart Scan
n
Scan on clustered table
n
Scan on Index Organized Table.
n
Fast Full Scan on Compressed Index
n
Fast Full scan on a reverse key index
n
More than 255 columns are referenced by query.
n
Optimized wants to scan to return rows in ROWID
order.
