In-Memory Checkpoint Services for Distributed Transaction Failover
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current server networks face challenges in maintaining high availability and rapid failover response due to the need for shared storage in checkpointing processes, which can lead to prolonged downtime during failures.
Innovation Solution
Implementing in-memory checkpoint services as a callable resource within a distributed two-phase transaction, utilizing an XA interface to manage and coordinate resource managers, allowing checkpoint data to be captured and restored without relying on common storage, enabling seamless failover and rollback processes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If shared storage is used for checkpointing, then checkpoint data can be accessed by multiple servers, but downtime during failover increases
Solution Approach 1:
The patent extracts the checkpointing function from shared storage dependencies and implements it as an in-memory resource manager within the distributed transaction framework. This allows checkpoint data to be stored and accessed without requiring shared storage infrastructure, thereby reducing failover downtime while maintaining availability.
Solution Approach 2:
The patent introduces an in-memory checkpoint resource manager as an intermediary between the application and the failover process. This intermediary manages checkpoint data in memory during transactions and coordinates its transfer during failover, eliminating the need for shared storage and reducing downtime.
2Reliability
If shared storage is required for checkpointing, then checkpoint data can be stored and accessed, but device complexity increases
Solution Approach 1:
The patent removes the shared storage requirement from the checkpointing mechanism by implementing in-memory checkpointing as a standalone resource manager. This extraction eliminates the need for complex shared storage configuration while maintaining fault tolerance capabilities.
Solution Approach 2:
The patent replaces the mechanical/shared storage system with an in-memory resource management system. Instead of relying on physical shared storage infrastructure, the system uses memory resources within the distributed transaction framework to store and manage checkpoint data, thereby reducing device complexity.
Data Source
AI summary
A method and apparatus for performing in-memory checkpoint services as a callable resource within a distributed transaction. As such, in-memory checkpoint processes can be utilized by an application as the application would use any resource available to the computer network via a distributed transaction.