Memory problems can occur when there is not enough free memory available to accommodate peak processing without paging and swapping. When free memory is exhausted, the entire process is paged or swapped to disk. Since disk access is much slower than memory access, performance suffers when this occurs.
The amount of memory needed depends on the program running, the language of that program, the DBMS accessed, the operating system, and the number of users logged on. The more memory available, the faster the system will run.
Memory usage is considered optimized when the hit ratio is 90 percent or better. The hit ratio is the ratio of finding rollback, table, index, and cluster data in memory rather than having to perform I/O from disk. The goal is to achieve as close to 100 percent utilization as possible during peak processing times.
The key point here is that more available memory translates into reduced requirements for swapping, thus improving overall system throughput.
These benefits are especially notable when processing a large model.
To avoid memory problems, especially when processing a large model:
|
Copyright © 2013 CA.
All rights reserved.
|
|