ASM Intelligent Data Placement

It is known that access to the data furthest from the spindle of a disk can be as much as 50% faster than for the data closest to the spindle. So placing frequently accessed data in the disk region furthest from the spindle will increase performance. This region is called the hot region and the region closest to the spindle, the cold region.

Some customers would never use the cold region of the disk nearest the spindle due to the increased latency. With Intelligent Data Placement, you can assign an ASM file to a disk region for best performance and use the lower performance region for less frequently accessed files. Using the disk region settings, you can ensure that frequently accessed data is placed on the outermost tracks (hot region). In addition, when files with similar access patterns are located physically close together, latency is reduced. Intelligent Data Placement also allows the placement of primary and mirror extents in different hot or cold regions.

Intelligent Data Placement settings can be specified for a file or in disk group templates. The disk region settings for a file can be modified after the file has been created. The disk region setting can improve I/O performance by placing more frequently accessed data in regions furthest from the spindle, while reducing your cost by increasing the usable space on a disk.

Guidelines for Intelligent Data Placement

Intelligent Data Placement works best for:

• Data files that are accessed at different rates
• Disk groups more than 50% full
• Disks with the geometry exposed to ASM (JBOD)

Intelligent Data Placement requires disk group attributes.

• COMPATIBLE.ASM = 11.2.0 or higher
• COMPATIBLE.RDBMS = 11.2.0 or higher

Assigning Files to Disk Regions

Create a file template:

ALTER DISKGROUP data ADD TEMPLATE datafile_hot

ATTRIBUTES ( HOT MIRRORHOT);

Alter the file attributes:

ALTER DISKGROUP data MODIFY FILE '+data/orcl/datafile/users.259.679156903'

ATTRIBUTES (COLD MIRRORCOLD);



Assigning Files to Disk Regions with Enterprise Manager






Monitoring Intelligent Data Placement

Performance benefit is dependent on:

• File access pattern
• Correct file access evaluation

Use performance views to select the candidate files:

• V$ASM_FILE
• V$ASM_DISK
• V$ASM_DISK_STAT

Performance benefit can be verified by testing.

• Review Automatic Workload Repository (AWR) reports before and after making a change.
• Use V$FILEMETRIC and V$FILEMETRIC_HISTORY.