Backup Coordinator for Distributed Transaction Failover

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Distributed transaction systems face delays and complexity due to the reliance on a single coordinator node, which can cause significant delays when it fails, as other participants must wait for the coordinator to recover before completing the transaction.

Innovation Solution

Implementing a backup coordinator that assumes control of the distributed transaction seamlessly in case of primary coordinator failure, using a prepare message with an address of the backup coordinator and exchanging heartbeat messages to ensure smooth failover, allowing the transaction to continue without interruption.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If a single coordinator node is used to manage distributed transactions, then the system maintains simplicity in coordination logic, but significant delays occur when the coordinator fails and participants must wait for recovery

Engineering Contradiction:

Improvetransaction completion reliabilityVSAvoidtransaction delay

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The patent designates a backup coordinator in advance before any failure occurs. The backup coordinator is pre-configured with the necessary transaction information and is ready to immediately take over if the primary coordinator fails, eliminating the need to wait for coordinator recovery and thus reducing transaction delays while maintaining system reliability

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The backup coordinator acts as an intermediary that can step in when the primary coordinator is unavailable. Participants can directly communicate with the backup coordinator using the same protocol messages, allowing transaction coordination to continue without interruption and preventing time loss due to coordinator failure

Inventive Principle:

Principle #24Intermediary (Mediator)

2Loss of time

If standby coordinators are implemented to minimize delays, then transaction continuity is improved, but system complexity increases due to clustering protocols and group communication requirements

Engineering Contradiction:

Improvecoordinator failure delayVSAvoidsystem protocol complexity

Core Design Contradiction:

Loss of timeVSDevice complexity

Solution Approach 1:

The backup coordinator is designed to handle the same transaction coordination tasks as the primary coordinator using the existing two-phase commit protocol. The same message types (prepare, commit, abort) are used for communication between participants and either coordinator, allowing the backup coordinator to seamlessly take over without requiring participants to learn new protocols or communication patterns, thus minimizing added complexity

Inventive Principle:

Principle #6Universality (Multi-functionality)

Solution Approach 2:

The backup coordinator maintains a copy of the transaction log and coordinator state, allowing it to assume the primary coordinator's role immediately upon failure. This copying approach simplifies the failover mechanism compared to complex clustering protocols, as the backup coordinator already has the necessary transaction information locally available without requiring complex group communication or distributed consensus mechanisms

Inventive Principle:

Principle #26Copying

Data Source

PatentUS7900085B2Backup coordinator for distributed transactions

Publication Date: 2011.03.01 RED HAT INC

AI summary

A primary coordinator generates a prepare message for a two-phase commit distributed transaction, the prepare message including an address of a backup coordinator. The primary coordinator maintains a transaction log of the distributed transaction, wherein the transaction log is accessible to both the primary coordinator and the backup coordinator. The prepare message is sent to a plurality of participants. The primary coordinator fails over to the backup coordinator without interrupting the distributed transaction.