Dynamically deploying execution nodes using system throughput

Patent metadata

Field	Value
Publication number	US-12481644-B2
Application number	US-202017014467-A
Country	US
Kind code	B2
Filing date	Sep 8, 2020
Priority date	Sep 8, 2020
Publication date	Nov 25, 2025
Grant date	Nov 25, 2025

Abstract

A method, system, and computer program product for using logical segments to help manage database tasks and performance for SQL parallelism. The method may include identifying a database object within a database resource of a database. The method may further include determining a database access method for the database object. The method may further include dividing data within the database object into a plurality of logical segments based on the database access method. The method may further include receiving a query for the database. The method may further include identifying the database object that corresponds to the query. The method may further include analyzing the logical segments of the database object. The method may further include distributing access to the database resource based on the logical segments of the database object.

First claim

What is claimed is:

1 . A computer-implemented method comprising:
dividing data within a database object within a database resource of a database into a plurality of logical segments based on an index scan comprising index only access and restrictions to database object access, wherein the dividing the data is based on a key-value from a non-leaf page, wherein a first logical segment of the plurality of logical segments comprises a range of data pages that are managed by a space map page, and wherein the space map page is a page that exists in both a table space and an index; identifying the database object that corresponds to a query for the database; analyzing the plurality of logical segments of the database object; and dynamically allocating access to the database resource based on the plurality of logical segments of the database object by:
creating a logical segment task map for the database object, wherein the logical segment task map records respective logical segments for the database object, tasks that are scheduled to process the respective logical segments, and indications when the respective logical segments have been processed;
tracking, using the logical segment task map, the tasks for the database object, their corresponding logical segments, and their corresponding processing indications;
in response to determining that the first logical segment from the plurality of logical segments is currently being processed:
waiting for the first logical segment to finish processing;
stopping subsequent second logical segments from being processed;
granting the query for the database resource; and
updating the logical segment task map; and

in response to determining that the first logical segment is not being processed:
processing the query; and
updating the logical segment task map.

2 . The method of claim 1 , wherein the dividing the data within the database object into the plurality of logical segments is further based on a join method and a join sequence.

3 . The method of claim 2 , wherein the join method is a nest loop join.

4 . The method of claim 2 , wherein the join sequence comprises at least a leading table and a non-leading table.

5 . A system comprising:
one or more computer processors; and one or more computer-readable storage media collectively storing program instructions, which, when executed by the one or more computer processors, are configured to cause the one or more computer processors to perform computer operations comprising: divide data within a database object within a database resource of a database into a plurality of logical segments based on an index scan comprising index only access and restrictions to database object access, wherein the dividing the data is based on a key-value from a non-leaf page, wherein a first logical segment of the plurality of logical segments comprises a range of data pages that are managed by a space map page, and wherein the space map page is a page that exists in both a table space and an index; identify the database object that corresponds to a query for the database; analyze the plurality of logical segments of the database object; and dynamically allocate access to the database resource based on the plurality of logical segments of the database object by:
creating a logical segment task map for the database object, wherein the logical segment task map records respective logical segments for the database object, tasks that are scheduled to process the respective logical segments, and indications when the respective logical segments have been processed;
tracking, using the logical segment task map, the tasks for the database object, their corresponding logical segments, and their corresponding processing indications;
in response to determining that the first logical segment from the plurality of logical segments is currently being processed:
waiting for the first logical segment to finish processing;
stopping subsequent second logical segments from being processed;
granting the query for the database resource; and
updating the logical segment task map; and

in response to determining that the first logical segment is not being processed:
processing the query; and
updating the logical segment task map.

6 . The system of claim 5 , wherein the dividing the data within the database object into the plurality of logical segments is further based on a join method and a join sequence.

7 . A computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable by a server to cause the server to perform a method, the method comprising:
dividing data within a database object within a database resource of a database into a plurality of logical segments based on an index scan comprising index only access and restrictions to database object access, wherein the dividing the data is based on a key-value from a non-leaf page, wherein a first logical segment of the plurality of logical segments comprises a range of data pages that are managed by a space map page, and wherein the space map page is a page that exists in both a table space and an index; identifying the database object that corresponds to a query for the database; analyzing the plurality of logical segments of the database object; and dynamically allocating access to the database resource based on the plurality of logical segments of the database object by:
creating a logical segment task map for the database object, wherein the logical segment task map records respective logical segments for the database object, tasks that are scheduled to process the respective logical segments, and indications when the respective logical segments have been processed;
tracking, using the logical segment task map, the tasks for the database object, their corresponding logical segments, and their corresponding processing indications;
in response to determining that the first logical segment from the plurality of logical segments is currently being processed:
waiting for the first logical segment to finish processing;
stopping subsequent second logical segments from being processed;
granting the query for the database resource; and
updating the logical segment task map; and

in response to determining that the first logical segment is not being processed:
processing the query; and
updating the logical segment task map.

8 . The method of claim 2 , wherein the join method is a sort merge join.

9 . The system of claim 6 , wherein the join method is a nest loop join.

10 . The system of claim 6 , wherein the join method is a sort merge join.

11 . The system of claim 6 , wherein the join sequence comprises at least a leading table and a non-leading table.

12 . The computer program product of claim 7 , wherein the dividing the data within the database object into the plurality of logical segments is further based on a join method and a join sequence.

13 . The computer program product of claim 12 , wherein the join method is a nest loop join.

14 . The computer program product of claim 12 , wherein the join sequence comprises at least a leading table and a non-leading table.

15 . The computer program product of claim 12 , wherein the join method is a sort merge join.

Assignees

• IBM

Inventors

• Li Shuo
• Wang Xiaobo
• Liu Xin Peng
• Wang Yue

Classifications

• G06F9/4843

  by program, e.g. task dispatcher, supervisor, operating system · CPC title

• G06F16/24544

  Join order optimisation · CPC title

• G06F16/2456

  Join operations · CPC title

• G06F16/2379 Primary

  Updates performed during online database operations; commit processing · CPC title

• G06F16/24532 Primary

  of parallel queries · CPC title