Adaptive Locking for Distributed Database Buffer Pool Contention
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In distributed database systems, existing approaches face challenges in reducing contention and avoiding deadlocks, particularly in shared disk cluster environments where physical locks are retained across nodes, leading to suboptimal performance due to complex deadlock detection and costly repeated acquisition cycles.
Innovation Solution
The implementation of adaptive locking mechanisms that identify locking priorities for buffer pools, determine lock requests based on these priorities, and grant locks accordingly, allowing clients to control locking at the node level and adapt to inter-node contention, thereby minimizing deadlocks and optimizing resource access.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If physical locks are retained across nodes in shared disk cluster environments, then resource access efficiency is improved, but deadlock detection complexity and contention increase
Solution Approach 1:
The patent segments the buffer pool into multiple parts or regions, allowing different locking strategies to be applied to different segments. This reduces the scope of locks that need to be coordinated across nodes, thereby reducing deadlock detection complexity while maintaining resource access efficiency in each segment
Solution Approach 2:
The patent implements dynamic locking strategies where the locking behavior adapts based on current system conditions, contention levels, and access patterns. This allows the system to optimize between retaining locks for efficiency and releasing locks to reduce contention and deadlock complexity, depending on real-time requirements
2Productivity
If physical locks are retained across nodes, then resource access efficiency is improved, but contention and deadlocks increase
Solution Approach 1:
The patent applies locking only to the necessary portions of resources rather than entire buffer pools. By using partial locking on specific segments or pages, the system maintains efficiency for accessed resources while reducing contention on unaccessed portions, thereby reducing deadlocks without sacrificing resource access efficiency
Solution Approach 2:
The patent implements feedback mechanisms that monitor contention levels and deadlock occurrences, dynamically adjusting locking strategies. When contention or deadlocks are detected, the system responds by modifying lock retention behavior, releasing locks strategically, or adjusting locking granularity to prevent future deadlocks while maintaining efficiency
3Reliability
If repeated lock acquisition cycles are performed, then resource access is maintained, but system performance degrades
Solution Approach 1:
The patent performs preliminary lock acquisition attempts and prepares alternative access strategies in advance. By anticipating potential lock unavailability and having pre-planned fallback mechanisms, the system reduces the need for repeated acquisition cycles, thereby maintaining resource access continuity while improving system performance through fewer retry operations
Data Source
AI summary
System, method, computer program product embodiments and combinations and sub-combinations thereof for adaptive locking of retained resources in a distributed database processing environment are provided. An embodiment includes identifying a locking priority for at least a portion of a buffer pool, determining lock requests based upon the identified locking priority, and granting locks for the lock requests.