Data Caches
Go Up to Supported Objects
Data Caches are devices stored in memory. If configured properly, they can substantially boost performance by allowing users to read or write data in memory instead of to or from physical disk. With System 11, you can now “borrow” memory from the default data cache and allocate it to customized data caches.
Named data caches let you achieve substantially greater concurrency. Data caches also increase the likelihood that Sybase ASE reads data from memory and not disk. DBArtisan lets you bind particularly active databases (for example, tempdb), tables (for example, syslogs and sysindexes) or indexes to dedicated data caches. In addition, you can boost performance by creating memory pools, each configured for different I/O throughput (2 K to 16 K), within every named data cache.
In spite of their potential, named data caches are not a panacea for performance problems. You should still carefully optimize your queries and indexing strategy. Poor configuration of named data caches can lead to worse performance because precious memory resources are misallocated. To determine data cache performance, you should examine the cache hit ratio.
Creating and editing
DBMS platform availability and object actions/operations supported
The table below lists the platforms on which this object type is available. It also lists the object actions available against objects of this type For an introduction to object actions and details on usage of specific actions, see Object actions.
| SYB ASE | |
|---|---|
|
✓ | |
|
✓ | |
|
✓ | |
|
✓ |