Deferred Lock Enforcement for Database Concurrency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing lock management systems in data management systems face inefficiencies due to computationally expensive and slow lock upgrades, particularly when handling multiple transactions and data resources, as they often require serial or sequential lock management, which can hinder concurrency and processing speed.
Innovation Solution
The implementation of a deferred enforcement lock system that dynamically interprets locks based on the transaction phase, allowing for parallel management of locks across multiple data resources, thereby omitting unnecessary lock upgrades and enhancing processing speed and reducing computational expense.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional lock management systems are used with lock upgrades, then data resource access control is ensured, but processing speed and computational efficiency deteriorate due to serial lock upgrading
Solution Approach 1:
The patent segments the lock management process into distinct phases: read phase with shared locks and commit phase with exclusive locks. This segmentation allows different lock types to be used at different stages, enabling parallel processing during the read phase while maintaining data integrity through exclusive locks during the commit phase, thus resolving the contradiction between reliability and productivity.
Solution Approach 2:
The patent implements dynamic lock type conversion where locks transition from shared locks during the read phase to exclusive locks during the commit phase. This dynamic adjustment of lock characteristics based on transaction phase allows the system to optimize for concurrency during reading while ensuring data integrity during commitment, thereby improving processing speed without sacrificing reliability.
2Manufacturing precision
If serial lock upgrading is implemented for multiple data resources, then lock management accuracy is maintained, but time consumption and processing efficiency worsen
Solution Approach 1:
The patent divides lock management into phase-specific operations: shared locks are acquired and held during the read phase for multiple data resources in parallel, then exclusive locks are acquired during the commit phase. This segmentation eliminates serial lock upgrading while maintaining accuracy through phase-based control, significantly reducing time consumption.
Solution Approach 2:
The patent performs preliminary lock acquisition of shared locks during the read phase before the commit phase begins. This preliminary action allows multiple data resources to be locked with shared locks in parallel, preparing the system for the subsequent exclusive lock acquisition during commit, thereby reducing overall time consumption while maintaining lock management accuracy.
3Productivity
If multiple locks are assigned to a transaction for parallel data resource access, then concurrency is improved, but lock management complexity and computational cost increase
Solution Approach 1:
The patent applies different lock qualities locally based on transaction phase: shared locks are used during the read phase to allow high concurrency with simple management, while exclusive locks are used during the commit phase to ensure data integrity. This local differentiation of lock characteristics reduces overall lock management complexity while maintaining high concurrency during the read phase.
Solution Approach 2:
The patent implements dynamic lock type conversion from shared to exclusive locks based on transaction phase transitions. This dynamic approach automatically adjusts lock characteristics without requiring complex manual management, allowing the system to maintain high concurrency during reading while ensuring data integrity during commitment, thereby improving productivity without proportionally increasing complexity.
Data Source
AI summary
Systems and methods for providing deferred lock enforcement for transactions are described. The method includes receiving a first request for access to a data resource for a transaction, assigning a first lock to the transaction, and receiving, from a candidate reader, a second request for access to the data resource. The method further includes determining an interpretation of the first lock as i) a first lock type responsive to determining that the transaction is in a read phase or ii) a second lock type responsive to determining that the transaction is in a commit processing phase, and assigning a second lock to the candidate reader responsive to interpreting the first lock as the first lock type, or declining to assign the second lock to the candidate reader responsive to interpreting the first lock as the second lock type.