Buffer pools are memory areas configured for a database. These areas temporarily store data and index pages from the database. If an application needs to read a data or index page and that page is already in the buffer pool, the application reads the page from the buffer pool¾a process that is faster than retrieving the page from disk. Avoiding disk I/O is the main issue when you are trying to improve database performance. Consequently, proper configuration of buffer pools is probably the most important consideration for performance tuning.
The buffer pool hit rate ratio indicates the percentage of page requests that were satisfied by pages already in the buffer pool. In other words, the hit rate represents the percentage of time that the database manager did not need to load a page from disk to fulfill a page request. When the high hit rate is high, the frequency of disk I/O is generally low, indicating that the buffer pools are performing effectively.
The Buffer Pool graph contains a single series that plots the Hit Rate for all data and index buffer pools configured for the database. The Hit Rate is the ratio of physical read requests to logical read requests for both data and index pages, expressed as a percentage.
The graph title area shows the number of index and data pages inserted into and retrieved from buffer pools during the last monitoring interval.
On this drilldown the following statistical information is presented:
The I/O Activity tab contains additional buffer pool details.
The Statistics tab contains entries for the buffer pool statistical counts used to generate this graph. The names for these statistics begin with Buffer Pool. These statistics are derived from the POOL data elements that the DB2 snapshot monitor captures.