Reducing database maintenance effort

Patent metadata

Field	Value
Publication number	US-12019598-B2
Application number	US-202016752101-A
Country	US
Kind code	B2
Filing date	Jan 24, 2020
Priority date	Jan 24, 2020
Publication date	Jun 25, 2024
Grant date	Jun 25, 2024

Abstract

Methods and systems for reducing database maintenance effort are disclosed. A method includes: collecting, by a computing device, a history of statistics for database objects; predicting, by the computing device, statistics using the history of statistics; testing, by the computing device, a workload using the predicted statistics; determining, by the computing device, at least one database object to maintain based on the testing the workload; and maintaining, by the computing device, the determined at least one database object.

First claim

What is claimed is:

1. A method comprising:
collecting, by a computing device, a history of statistics for database objects;
predicting, by the computing device, a first maintenance window based on statistics for the database objects using the history of statistics for the database objects;
testing, by the computing device, a workload using the predicted statistics for the database objects, comprising:
saving a current access path of the database object, the current access path having a first processing cost;
generating an updated access path of the database object using the predicted statistics, the updated access path having a reduced processing cost; and
validating the updated access path of the database object,
wherein the reduced processing cost comprises a reduction from the first processing cost based on the updated access path of the database object;

determining, by the computing device, at least one database object to maintain based on the testing the workload; and
maintaining, by the computing device, the determined at least one database object by performing at least one database maintenance task during a reduced maintenance window, wherein the reduced maintenance window comprises a reduction of the first maintenance window based on the determining the at least one database object to maintain.

2. The method according to claim 1 , wherein the testing the workload is performed offline.

3. The method according to claim 1 , wherein the determined at least one database object is determined based on the validating the updated access paths.

4. The method according to claim 1 , wherein, for each database object of the determined at least one database object, maintaining the database object further comprises switching the current access path of the database object.

5. The method according to claim 1 , wherein the predicted statistics correspond to a next maintenance window of a relational database.

6. The method according to claim 1 , wherein the predicting the statistics is performed using a prediction model, and
further comprising modifying the prediction model based on feedback received regarding the predicted statistics.

7. A computer program product, the computer program product comprising one or more computer readable storage media having program instructions collectively stored on the one or more computer readable storage media, the program instructions executable to:
collect a history of statistics for database objects;
predict statistics for the database objects corresponding to an upcoming database maintenance window using the history of statistics for the database objects;
determine an updated access path of a database object using the predicted statistics; and
switch an access path of the database object to the updated access path at a reduced database maintenance window.

8. The computer program product according to claim 7 , wherein the predicting the statistics comprises using a line regression method.

9. The computer program product according to claim 7 , wherein the predicting the statistics comprises using a random forest regression method.

10. The computer program product according to claim 7 , wherein the predicted statistics further comprise a fullkeycard statistic, and
wherein the database objects comprise representations of information in a form as used in an object-oriented or object-oriented-compatible programming language.

11. The computer program product according to claim 7 , wherein the switching the access path of the database object is performed in response to the updated access path being validated.

12. The computer program product according to claim 7 , the program instructions further being executable to compare newly collected statistics with the predicted statistics to determine an accuracy of the predicted statistics.

13. A system comprising:
a hardware processor, a computer readable memory, and a computer readable storage medium associated with a computing device;
program instructions to collect a history of statistics for database objects;
program instructions to predict a first maintenance window based on statistics for the database objects using the history of statistics for the database objects, wherein the program instructions to predict the first maintenance window based on statistics for the database objects using the history of statistics for the database objects comprise program instructions to determine whether the database objects comprise constant growth tables or volatile tables;
program instructions to test a workload using the predicted statistics for the database objects, the test comprising:
saving a current access path of the database object, the current access path having a first processing cost;
generating an updated access path of the database objects using the predicted statistics, the updated access path having a reduced processing cost; and
validating the updated access path of the database object,
wherein the reduced processing cost comprises a reduction from the first processing cost based on the updated access path of the database object;

program instructions to determine at least one database object to maintain based on the testing the workload; and
program instructions to maintain the determined at least one database object using at least one database maintenance task during a reduced maintenance window.

14. The system according to claim 13 , wherein the testing the workload is performed offline.

15. The system according to claim 13 , wherein the at least one database object is determined based on the validating the updated access paths.

16. The system according to claim 13 , wherein, for each database object of the determined at least one database object, maintaining the database object further comprises switching the current access path of the database object to the updated access path of the database object.

17. The system according to claim 13 , wherein the predicting the statistics is performed using a prediction model, and
further comprising modifying the prediction model based on feedback received regarding the predicted statistics.

18. The method according to claim 1 , wherein predicting the statistics for the database objects using the history of statistics for the database objects comprises:
predicting a cluster ratio statistic for the database objects using the history of statistics for the database objects.

Assignees

• IBM

Inventors

• Li Shuo
• Sun Shengyan
• Wang Xiaobo
• Jiang peng hui

Classifications

• G06F11/3452

  Performance evaluation by statistical analysis · CPC title

• G06N5/02

  Knowledge representation; Symbolic representation · CPC title

• G06N5/04

  Inference or reasoning models · CPC title

• G06N20/20

  Ensemble learning · CPC title

• G06F11/3428

  Benchmarking · CPC title