Click here to view and discuss this page in DocCommentXchange. In the future, you will be sent there automatically.

SQL Anywhere 10.0.1 » SQL Anywhere Server - SQL Usage » Query Optimization and Execution » Physical data organization and access

Disk allocation for inserted rows Next Page

Table and page sizes

The page size you choose for your database can affect the performance of your database. In general, smaller page sizes are likely to benefit operations that retrieve a relatively small number of rows from random locations. By contrast, larger pages tend to benefit queries that perform sequential table scans, particularly when the rows are stored on pages in the order the rows are retrieved via an index. In this situation, reading one page into memory to obtain the values of one row may have the side effect of loading the contents of the next few rows into memory. Often, the physical design of disks permits them to retrieve fewer large blocks more efficiently than many small ones.

For each table, SQL Anywhere creates a bitmap that reflects the position of each table page in the entire dbspace file. The database server uses the bitmap to read large blocks (64 KB) of table pages, instead of single pages at a time. This efficiency, also known as group reads, reduces the total number of I/O operations to disk, and improves performance. Users cannot control the database server's criteria for bitmap creation or usage.

Should you choose a larger page size, such as 8 KB, you may want to increase the size of the cache because fewer large pages can fit into a cache of the same size. For example, 1 MB of memory can hold 512 pages that are each 2 KB in size, but only 128 pages that are 8 KB in size. Determining the proper page ratio of page size to cache size depends on your database and the nature of the queries your application performs. You can conduct performance tests with various cache sizes. If your cache cannot hold enough pages, performance suffers as the database server begins swapping frequently-used pages to disk. This is particularly important when using SQL Anywhere on a Windows CE device, since larger page sizes may have a greater amount of internal fragmentation.

SQL Anywhere attempts to fill pages as much as possible. Empty space accumulates only when new objects are too large to fit empty space on existing pages. Consequently, adjusting the page size may not significantly affect the overall size of your database.

Page size also affects indexes. Each index lookup requires one page read for each of the levels of the index plus one page read for the table page, and a single query can require several thousand index lookups. Page size can significantly affect fan-out, in turn affecting the depth of index required for a table. A large fan-out often means that fewer index levels are required, which can improve searches considerably. For large databases that have tables with a significant numbers of rows, 8 KB pages may be warranted for the best performance. It is strongly recommended that you test performance (as well as other behavior aspects) when choosing a page size. Then, choose the smallest page size that gives satisfactory results. It is particularly important to pick the correct and reasonable page size if more than one database will be started on the same server.

See also