Remote Mapping Resolution for Sliced Logical Address Read I/O
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data storage systems face inefficiencies in cache usage due to cache misses and traversal of metadata pages, leading to increased performance penalties and latency in mapping logical to physical addresses, especially in systems with multiple processing nodes.
Innovation Solution
A method of partitioning ownership of virtual layer blocks (VLBs) among multiple nodes, with exclusive ownership assigned to each node, and utilizing remote procedure calls (RPCs) for nodes to perform mapping resolution processing, allowing efficient caching and reduced traversal of metadata pages.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If multiple nodes cache metadata pages to improve access speed, then cache hit rate increases, but duplicate caching across nodes increases memory usage and reduces cache efficiency
Solution Approach 1:
The patent segments the metadata cache into node-specific portions, where each node caches only the metadata pages it owns or is responsible for. This segmentation prevents duplicate caching across nodes while ensuring each node has fast access to its assigned metadata, resolving the contradiction between access speed and memory efficiency.
Solution Approach 2:
Each node is assigned specific ownership of virtual layer blocks and corresponding metadata pages, creating local quality differences in the cache system. Nodes cache metadata locally based on their ownership responsibilities rather than universally caching all metadata, optimizing both access speed for owned metadata and overall cache utilization across the distributed system.
2Measurement precision
If nodes traverse metadata pages to resolve logical to physical address mapping, then address mapping is achieved, but traversal latency increases I/O operation time
Solution Approach 1:
The system performs preliminary actions by pre-establishing ownership assignments of virtual layer blocks to specific nodes and pre-caching the corresponding metadata pages at those nodes. When an I/O operation requires address mapping, the owning node already has the necessary metadata cached, eliminating the need for time-consuming traversal and reducing I/O latency while maintaining accurate mapping.
3Productivity
If exclusive ownership of virtual layer blocks is assigned to specific nodes, then cache efficiency improves, but system complexity increases due to ownership management
Solution Approach 1:
Each node autonomously manages its own cache based on its assigned ownership of virtual layer blocks. Nodes independently determine which metadata pages to cache based on their ownership responsibilities without requiring complex centralized coordination or communication with other nodes. This self-service approach improves cache efficiency while minimizing the complexity of ownership management.
Data Source
AI summary
Logical address space portions and virtual layer blocks (VLBs) can be partitioned into multiple sets. Each of multiple nodes in a system can be assigned exclusive ownership of one of the multiple sets. In at least one embodiment, for a read I/O which is received at a first node and directed to a logical address LA1 that is owned by a second node, the first node can request that the second owning node perform resolution processing for LA1. The second node can return either a VLB address or a PLB address based on whether the second node owns a VLB used in mapping LA1 to a corresponding physical location PA1 which includes content C1 stored at LA1. The second node can set a flag in its response to indicate whether a returned address is a VLB address or a PLB address.