Distributed Transaction Processing via Centralized Commit Log Coordination

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

In distributed transaction processing, the accuracy of transaction processing cannot be ensured when established processing steps are interrupted, leading to inconsistent transaction statuses across distributed devices.

Innovation Solution

A distributed transaction processing method that involves writing a commit log to a second object database set within a target database cluster, allowing for consistent resumption of transaction processing even if local caching fails.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Speed

If a distributed transaction uses local caching for commit logs, then processing speed is improved, but reliability deteriorates when catastrophic failures occur during commit processes

Engineering Contradiction:
Improvetransaction processing speedVSAvoidtransaction commit reliability
Core Design Contradiction:
SpeedVSReliability

Solution Approach 1:

The patent introduces a coordinator node as an intermediary between distributed devices to manage commit logs and coordinate transaction commits. The coordinator collects commit logs from all distributed devices, ensures consistent ordering, and manages the commit process centrally, thereby maintaining reliability while enabling distributed processing.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent implements preliminary actions by requiring all distributed devices to pre-commit their local changes and record commit logs before the final commit is confirmed. This preliminary commitment phase allows the system to detect potential failures early and rollback if necessary, ensuring atomicity across the distributed system.

Inventive Principle:
Principle #10Preliminary action

2Manufacturing precision

If distributed transaction processing follows established steps, then processing accuracy is maintained, but device complexity increases due to coordination requirements across multiple distributed devices

Engineering Contradiction:
Improvetransaction processing accuracyVSAvoiddistributed system complexity
Core Design Contradiction:
Manufacturing precisionVSDevice complexity

Solution Approach 1:

The patent segments the distributed transaction system into distinct functional components: coordinator nodes that manage commit coordination, participant nodes that execute local transactions, and a distributed commit log storage system. This segmentation allows each component to focus on specific tasks, maintaining processing accuracy while managing complexity through modular design.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The coordinator node performs multiple functions including collecting commit logs, determining commit order, managing rollback operations, and ensuring consistency across all distributed devices. This multi-functionality reduces the need for separate specialized components, thereby managing system complexity while maintaining transaction accuracy.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Data Source

PatentEP3812930B1Distributed transaction processing method and related device
Publication Date: 2025.04.23 TENCENT TECHNOLOGY (SHENZHEN) CO LTD
  • EP3812930B1 patent drawingFigure 1
  • EP3812930B1 patent drawingFigure 2
  • EP3812930B1 patent drawingFigure 3

AI summary

Disclosed are a distributed transaction processing method and a related device. The method comprises: obtaining a distributed transaction processing request, the distributed transaction processing request being used for requesting to execute a processing operation on account data in N accounts, the account data in the N accounts being located in M first object database sets comprised in a target database cluster; writing a commit log corresponding to the distributed transaction processing request to a second object database set, wherein the commit log records a commit indication result, and the second object database set is comprised in the target database cluster; after distributed transaction processing is restored, obtaining the commit log from the second object database set; and when the commit indication result indicates to commit the processing operation executed on the account data in the N accounts, executing the processing operation on the account data in the N accounts. The present invention solves the technical problem of not being able to ensure the accuracy of the distributed transaction processing when a given step of a distributed transaction is interrupted.