Agent flow arrangement management

Patent metadata

Field	Value
Publication number	US-11221870-B2
Application number	US-201916292987-A
Country	US
Kind code	B2
Filing date	Mar 5, 2019
Priority date	Apr 25, 2017
Publication date	Jan 11, 2022
Grant date	Jan 11, 2022

Abstract

Disclosed aspects relate to agent flow arrangement management in a distributed commit processing environment. A first set of agent utilization data may be collected with respect to a first commit processing agent. A second set of agent utilization data may be collected with respect to a second commit processing agent. An agent flow arrangement may be determined based on a first value with respect to the first set of agent utilization data exceeding a second value with respect to the second set of agent utilization data. The agent flow arrangement may have the first commit processing agent subsequent to the second commit processing agent. The distributed commit operation may be processed using the agent flow arrangement which has the first commit processing agent subsequent to the second commit processing agent.

First claim

What is claimed is:

1. A computer-implemented method for agent flow arrangement management in a distributed two-phase commit processing environment, the method comprising:
collecting, in the distributed two-phase commit processing environment, a first set of agent utilization data with respect to a first commit processing agent;
collecting, in the distributed two-phase commit processing environment, a second set of agent utilization data with respect to a second commit processing agent, wherein the second commit processing agent is associated with a different transaction than a transaction associated with the first commit processing agent, and wherein the different transaction and the transaction associated with the first commit processing agent each include a plurality of work operations to complete a corresponding transaction;
determining, based on a first value with respect to the first set of agent utilization data exceeding a second value with respect to the second set of agent utilization data, an agent flow arrangement to process a distributed commit operation, wherein the agent flow arrangement has the first commit processing agent subsequent to the second commit processing agent;
processing, in the distributed two-phase commit processing environment, the distributed commit operation using the agent flow arrangement which has the first commit processing agent subsequent to the second commit processing agent; and
wherein determing the agent flow arrangement is based on,
quantities of volume of recovery log data written for processing agents with respect to a job unit,
quantities or distinct recoverable changes by the processing agents with respect to the job unit,
quantities of locks held by the processing agents with respect to the job unit, and
quantities of hardware processor usage by processing agents with respect to the job unit.

2. The method of claim 1 , further comprising:
configuring the distributed commit operation to include a prepare operation for the second commit processing agent but not the first commit processing agent.

3. The method of claim 1 , further comprising:
arranging, in the agent flow arrangement, the first commit processing agent to be a last agent to process the distributed commit operation.

4. The method of claim 1 , further comprising:
configuring the first set of agent utilization data to indicate a first quantity of work that has been performed by the first commit processing agent;
configuring the second set of agent utilization data to indicate a second quantity of work that has been performed by the second commit processing agent;
comparing the first quantity of work with the second quantity of work;
calculating that the first quantity of work exceeds the second quantity of work; and
determining, based on the first quantity of work exceeding the second quantity of work, the agent flow arrangement to process the distributed commit operation.

5. The method of claim 1 , further comprising:
configuring the first set of agent utilization data to indicate a first quantity of work that is expected to be performed by the first commit processing agent;
configuring the second set of agent utilization data to indicate a second quantity of work that is expected to be performed by second the commit processing agent;
comparing the first quantity of work with the second quantity of work;
calculating that the first quantity of work exceeds the second quantity of work; and
determining, based on the first quantity of work exceeding the second quantity of work, the agent flow arrangement to process the distributed commit operation.

6. The method of claim 1 , further comprising:
configuring the first set of agent utilization data to indicate a first volume of recovery log data written for the first commit processing agent with respect to a job unit;
configuring the second set of agent utilization data to indicate a second volume of recovery log data written for the second commit processing agent with respect to the job unit;
comparing the first volume of recovery log data written with the second volume of recovery log data written;
calculating that the first volume of recovery log data written exceeds the second volume of recovery log data written; and
determining, based on the first volume of recovery log data written exceeding the second volume of recovery log data written, the agent flow arrangement to process the distributed commit operation.

7. The method of claim 1 , further comprising:
configuring the first set of agent utilization data to indicate a first quantity of distinct recoverable changes by the first commit processing agent with respect to a job unit;
configuring the second set of agent utilization data to indicate a second quantity of distinct recoverable changes by the second commit processing agent with respect to the job unit;
comparing the first quantity of distinct recoverable changes with the second quantity of distinct recoverable changes;
calculating that the first quantity of distinct recoverable changes exceeds the second quantity of distinct recoverable changes; and
determining, based on the first quantity of distinct recoverable changes exceeding the second quantity of distinct recoverable changes, the agent flow arrangement to process the distributed commit operation.

8. The method of claim 1 , further comprising:
configuring the first set of agent utilization data to indicate a first quantity of locks held by the first commit processing agent with respect to a job unit;
configuring the second set of agent utilization data to indicate a second quantity of locks held by the second commit processing agent with respect to the job unit;
comparing the first quantity of locks held with the second quantity of locks held;
calculating that the first quantity of locks held exceeds the second quantity of locks held; and
determining, based on the first quantity of locks held exceeding the second quantity of locks held, the agent flow arrangement to process the distributed commit operation.

9. The method of claim 1 , further comprising:
configuring the first set of agent utilization data to indicate a first quantity of hardware processor usage by the first commit processing agent with respect to a job unit;
configuring the second set of agent utilization data to indicate a second quantity of hardware processor usage by the second commit processing agent with respect to the job unit;
comparing the first quantity of hardware processor usage with the second quantity of hardware processor usage;
calculating that the first quantity of hardware processor usage exceeds the second quantity of hardware processor usage; and
determining, based on the first quantity of hardware processor usage exceeding the second quantity of hardware processor usage, the agent flow arrangement to process the distributed commit operation.

10. The method of claim 1 , further comprising:
coordinating, by the first commit processing agent, the agent flow arrangement to process the distributed commit operation; and
selecting, by the first commit processing agent, the first commit processing agent to be a last agent to process the distributed commit operation.

11. The method of claim 1 , further comprising:
avoiding a prepare operation for the first commit processing agent.

12. The method of claim 1 , further comprising:
collecting, in the distributed two-phase commit processing environment, a third set of agent utilization data with respect to a third commit processing agent;
determining, based on the first value with respect to the first set of agent utilization data exceeding a third value with respe

Assignees

• IBM

Inventors

• Armitage Joshua H
• Clarke Michael P
• Kaputin John A W
• Kwan King-Yan
• Wright Andrew

Classifications

• G06F11/1471

  involving logging of persistent data for recovery · CPC title

• G06F9/466 Primary

  Transaction processing · CPC title

• G06F11/1474

  in transactions (updating of structured data in databases G06F16/23) · CPC title

• G06F2201/84

  Using snapshots, i.e. a logical point-in-time copy of the data · CPC title

• G06F11/3409

  for performance assessment · CPC title