Cluster Lock Manager Throughput via Validity Entry Aging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing cluster/lock processing systems are costly and inefficient due to reliance on specialized hardware and firmware, which limits their scalability and reliability, and struggle with managing redundancy and failure recovery without incurring high operational costs.
Innovation Solution
Implementing a cluster/lock processing system using commodity hardware and software, with redundant servers operating in a master/slave configuration under a standard operating system like Windows, and employing a cluster lock manager to manage locks, caching, and inter-host messaging, allowing for efficient recovery from failures without specialized processing needs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If specialized hardware and firmware are used in cluster/lock processing systems, then system reliability is improved, but system cost and complexity increase significantly
Solution Approach 1:
The patent creates a virtual copy of the lock manager functionality by implementing a shadow lock manager that maintains a duplicate copy of the lock database. This shadow copy can take over immediately upon failure of the primary lock manager, providing high availability without requiring specialized redundant hardware. The shadow lock manager is a software-based copy that mirrors the primary system's state.
Solution Approach 2:
The patent employs commodity hardware components that can be replaced easily and inexpensively compared to specialized hardware. The system uses standard off-the-shelf servers, storage devices, and network equipment rather than proprietary components, reducing both initial cost and replacement cost while maintaining system reliability through software-based redundancy.
2Ease of manufacture
If commodity hardware and software are used instead of specialized components, then system cost is reduced, but system reliability and performance may deteriorate
Solution Approach 1:
The patent implements proactive failure preparation by continuously maintaining a shadow lock manager that is pre-synchronized with the primary lock manager. When failure is detected, the system has already prepared a ready-to-takeover backup, cushioning against the reliability risks of using commodity hardware. This beforehand preparation eliminates the need for complex real-time synchronization during failure scenarios.
Solution Approach 2:
The patent introduces a communication interface layer that mediates between the lock manager and the cluster nodes. This intermediary layer abstracts the underlying hardware differences, allowing commodity hardware to be used while maintaining consistent system behavior. The communication interface handles protocol translation and data formatting, ensuring reliable operation across standard hardware platforms.
3Reliability
If redundant servers are implemented for failure recovery, then system availability is improved, but system complexity and operational costs increase
Solution Approach 1:
The patent merges the backup functionality into the existing cluster infrastructure by having regular cluster nodes serve as shadow lock managers. Instead of adding separate dedicated backup hardware, the system combines redundancy functions with existing components. The shadow lock manager resides on a cluster node and shares resources with other cluster functions, reducing overall system complexity.
Solution Approach 2:
The patent makes cluster nodes universal by enabling them to serve multiple functions: they can act as regular processing nodes, shadow lock managers, or failover targets. This multi-functionality eliminates the need for specialized single-purpose backup hardware. The same hardware platform performs diverse roles depending on system state and failure scenarios, reducing total component count and complexity.
4Reliability
If validity operations are performed frequently to maintain cache consistency, then data consistency is improved, but system throughput decreases due to processing overhead
Solution Approach 1:
The patent applies partial action by performing validity operations selectively rather than universally. The shadow lock manager performs validity checks and updates only for the specific data pages or segments that are affected by failures or modifications, rather than validating entire databases. This partial approach maintains necessary consistency while avoiding excessive processing overhead on unaffected data.
Solution Approach 2:
The patent performs preliminary validity operations during normal operation to prepare data for potential failover. The shadow lock manager continuously synchronizes with the primary lock manager and pre-validates data structures before failures occur. This preliminary preparation reduces the burden of validity operations during actual failure recovery, improving both consistency and throughput during critical moments.
Data Source
AI summary
An apparatus for and method of enhancing reliability and performance within a cluster lock processing system having a relatively large number of commodity instruction processors which are managed by a highly scalable, off the shelf platform. Because the commodity processors have virtually no system viability features such as memory protection, failure recovery, etc., the cluster/lock processors assume the responsibility for providing these functions. The low cost of the commodity instruction processors makes the system almost linearly scalable. The cluster/locking, caching, and mass storage accessing functions are fully integrated into a single hardware platform which performs the role of the cluster/lock master. The validity operation throughput of the clustered systems manager is increased by aging out validity entries for each of the process owners via a background operation. This minimizes the number of exclusive locks that must be utilized while performing a validity operation.