Binary Buddy Allocation for Root Indirect Block Lock Contention
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In hierarchical log structured storage systems, the shared root indirect block page becomes a performance bottleneck due to cache level locks, leading to measurable performance degradation as multiple CPUs access the same page, especially when multiple sockets are involved.
Innovation Solution
Implementing a binary buddy allocation scheme with a preferred alignment value to distribute volumes across different root indirect block pages, thereby reducing lock contention and improving performance by ensuring each volume is allocated to a distinct root page, and iteratively splitting extents to match the preferred alignment value.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If multiple volumes share the same root indirect block page, then memory utilization is improved, but lock contention increases and performance degrades
Solution Approach 1:
The patent segments the root indirect block pages by introducing an alignment-based allocation mechanism that divides the address space into distinct regions. Each volume is allocated to a specific root page based on alignment rules, effectively segmenting the previously shared resource into dedicated segments. This reduces lock contention while maintaining memory utilization efficiency.
Solution Approach 2:
The patent applies local quality by making each root indirect block page have a specific characteristic (dedicated to a particular volume based on alignment). Instead of uniform sharing, each page is optimized for local access patterns, reducing cache line movement and improving performance for specific volume operations.
2Adaptability or versatility
If multiple CPUs access the same root page, then resource sharing is improved, but cache line movement between sockets increases latency
Solution Approach 1:
By segmenting access to root pages through alignment-based allocation, the patent ensures that each CPU primarily accesses its own dedicated root page. This segmentation reduces the frequency of cache line movements between sockets while maintaining resource sharing capabilities through the structured allocation scheme.
Solution Approach 2:
The alignment value acts as an intermediary mechanism that mediates between resource sharing needs and latency reduction. It provides a structured way to allocate access rights to root pages, reducing direct contention and cache line bouncing between CPUs while still enabling efficient resource utilization.
3Ease of manufacture
If standard binary buddy allocation is used, then allocation simplicity is maintained, but root page sharing causes synchronization bottlenecks
Solution Approach 1:
The patent modifies the binary buddy allocation algorithm by introducing an alignment parameter (preferred alignment value). This parameter change transforms the allocation behavior to respect root page boundaries, reducing synchronization bottlenecks while maintaining the fundamental simplicity of the buddy allocation approach through iterative splitting and merging operations.
Data Source
AI summary
A method, computer program product, and computer system for setting a preferred alignment value to a size of an address space mapped by one or more root pages. An allocation request may be received for the address space. A binary buddy allocation scheme may be executed to allocate an extent for the allocation request based upon, at least in part, the preferred alignment value.