Lock-less Multidimensional Database Index Cache Concurrency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Multidimensional database systems face performance issues due to lock contention, which increases with the number of processing cores, leading to inefficiency and reduced concurrency.
Innovation Solution
Implementing lock-less algorithms and data structures using specific hardware-level instructions, such as compare-and-swap functions, to enable atomic operations and reduce contention, thereby improving concurrency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional locking mechanisms are used for data structure access, then data consistency is maintained, but concurrency is reduced and performance deteriorates with increased processing cores
Solution Approach 1:
The patent extracts the locking mechanism entirely from the data structure access protocol and replaces it with lock-less algorithms. The locking function is removed and replaced with atomic operations using hardware instructions like compare-and-swap, allowing concurrent access without traditional locks while maintaining data consistency through atomicity guarantees.
Solution Approach 2:
The patent substitutes the mechanical locking system with hardware-level atomic instructions. Instead of using software-based locks that require thread blocking and context switching, the system uses CPU atomic instructions (compare-and-swap, fetch-and-add) that provide the necessary synchronization through hardware enforcement, eliminating the need for traditional locking mechanisms.
2Power
If the number of processing cores is increased to improve performance, then processing capacity increases, but lock contention increases leading to reduced efficiency
Solution Approach 1:
The patent segments the shared data structure into multiple independently accessible regions or uses per-core data structures that can be merged atomically. This segmentation allows multiple cores to operate on different segments simultaneously without interfering with each other, reducing lock contention while maintaining data integrity through periodic atomic merge operations.
Solution Approach 2:
The patent introduces an intermediary atomic operation layer between core access and data structure modification. Hardware atomic instructions act as intermediaries that coordinate access without requiring traditional locks, allowing multiple cores to proceed concurrently while ensuring only one core can modify a given data element at a time, thus reducing idle time.
3Productivity
If lock-less algorithms are implemented, then concurrency is improved and scalability increases, but implementation complexity increases
Solution Approach 1:
The patent changes the fundamental parameter of synchronization from software-based locking to hardware-based atomic operations. By leveraging CPU atomic instructions that are natively supported by modern processors, the implementation complexity is reduced compared to software lock mechanisms, as the hardware provides the synchronization guarantees directly without requiring complex software protocols.
Data Source
AI summary
In accordance with an embodiment, described herein is a system and method for use of lock-less data structures and processes with a multidimensional database computing environment. Lock-less algorithms or processes can be implemented with specific hardware-level instructions so as to provide atomicity. A memory stores an index cache retaining a plurality of index pages of the multidimensional database. A hash table indexes index pages in the index cache, wherein the hash table is accessible by a plurality of threads in parallel through application of the lock-less process.