Distributed Lock Manager Active Lock Recovery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In networks, rebuilding lock information after a node failure is time-consuming due to the need to manage hundreds of thousands of locks, causing significant delays in application execution.
Innovation Solution
Implementing a status flag for locks, where active locks are rebuilt and inactive locks are discarded, reducing the time required to rebuild lock information by several orders of magnitude.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If all locks are rebuilt after a node failure, then complete lock information is restored, but the recovery time becomes excessively long
Solution Approach 1:
The patent extracts and identifies only the active locks from the complete lock set after a node failure. By using a lock status flag to distinguish active locks from inactive ones, the system recovers only the necessary subset of lock information, eliminating the need to rebuild all locks and thereby significantly reducing recovery time while maintaining reliability for active resources.
Solution Approach 2:
The patent applies partial action by rebuilding only the necessary portion of lock information (active locks) rather than performing a complete rebuild of all locks. This selective approach recovers sufficient lock data to maintain system functionality while avoiding the excessive time cost of processing inactive locks that do not require recovery.
2Loss of time
If lock status tracking is implemented, then recovery time is reduced, but system complexity increases
Solution Approach 1:
The lock manager automatically tracks and maintains the status flag for each lock without requiring external intervention or complex coordination mechanisms. Each lock's active/inactive state is self-managed through simple flag updates, allowing the system to quickly identify which locks need recovery after a failure without adding significant complexity to the overall lock management architecture.
Data Source
AI summary
A system is disclosed. A node may include a resource. A lock manager may include storage for a data structure for a lock. The data structure may include a first identifier for the resource, a second identifier for an application, and a status for the lock.