Granted Patent B2

US 8,341,181 · App. 12/908,917 · Granted Dec 25, 2012

Abstract

Claims (32)

1. A method for performance tuning a database running on a computing system, the computing system comprising at least one processor and a first random access memory, the database having data structures, attributes, and files, the method comprising:

setting a database-allowed core value attribute between 500 megabytes and half size of the first random access memory for the database, the database-allowed core value representing the amount of the first random access memory which will be available to the database;

allocating to the database a first portion of the first random access memory corresponding to the database-allowed core value;

setting a memory resident limit value between 0 and the database-allowed core value for the database;

declaring small and heavily accessed data structures within the database as memory resident structures and storing the small and heavily accessed structures within the first portion of the first random access memory up to the memory resident limit value;

analyzing the database to determine whether read-ahead should be enabled for the data structures, the analysis comprising a review of a ratio of logical to physical read-I/O's used by the database;

analyzing the database to determine an appropriate reblocking factor between 1 and 60, and applying the reblocking factor to the database; and

associating system buffers ranging from 1000 to 20000 with the data structures, the system buffers being stored in the first portion of the first random access memory.

2. The method of claim 1 , the database includes a Data Management System II (DMS II) database.

3. The method of claim 1 , wherein the size of the first random access memory is between 3 to 6 Gigabytes.

4. The method of claim 1 further comprising consolidating the data structures having unused space.

5. The method of claim 1 , wherein the system buffers include a set of default buffers having a value set to 1000 per data structure.

6. The method of claim 5 , wherein the system buffer includes a set of additional buffers assigned to data structures based on the size and access frequency of the data structures.

7. The method of claim 1 , wherein the memory resident limit is selected based on the size of memory resident structures and the memory size.

8. The method of claim 7 , wherein the memory resident limit is set to 510 megabytes.

9. The method of claim 1 further comprising setting a reorganization tasklimit between 1 and twice the number of processing units operating on the database reorganization process.

10. The method of claim 9 further comprising setting number of parallel executed tasks to between 1 and number of index sets for the reorganized data set.

11. The method of claim 10 , wherein the number of parallel executed tasks being set to 6.

12. The method of claim 9 further comprising setting the allowed core value for the reorganization process to 2.7 gigabytes.

13. The method of claim 12 , wherein setting the allowed core value and the reorganization tasklimit is based on:

a data structure population;

number of index sets for the data structure; or

number of available processing units.

14. The method of claim 13 , wherein the allowed core value is set to 2.7 gigabytes.

15. The method of claim 1 further comprising identifying the database sort process based on number of records in the database files.

16. The method of claim 15 further comprising setting the amount of memory and disk used for the database sort process based on a set of parameters.

17. The method of claim 16 , wherein the set of parameters include:

a record length;

a key length; or

the number of records.

18. The method of claim 16 , wherein the memory size for the sort process is set to 112.5 megabytes for a database file having 10 million records.

19. The method of claim 16 , wherein the disk size is set to 4050 megabytes.

Assignments (9)