Remote Index Node Lookup Using Hot-Spot Range Index Caching

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

In distributed storage systems, accessing target data in storage nodes involves significant network transmission, leading to decreased data access performance due to recursive access of index structures.

Innovation Solution

Implement a popularity-aware range index caching solution that caches frequently accessed hot-spot nodes in the compute node, using a cache module to temporarily store index nodes and manage cache using heat statistics and management modules to reduce remote access.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If recursive access of index structure is used to read target data, then data can be accurately retrieved, but network transmission increases and data access performance decreases

Engineering Contradiction:
Improvedata retrieval accuracyVSAvoiddata access performance
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

The patent pre-loads index nodes from the storage node into the compute node's cache before actual data access operations. This preliminary action stores frequently accessed index nodes (particularly hot-spot nodes at upper layers of the index tree) in local memory, so that when data access is needed, the compute node can directly access the cached index nodes without performing recursive network access to the storage node, thereby improving data access performance while maintaining retrieval accuracy

Inventive Principle:
Principle #10Preliminary action

2Productivity

If all index nodes are cached in the compute node, then data access performance improves, but memory usage and cache management complexity increase

Engineering Contradiction:
Improvedata access performanceVSAvoidcache management complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent implements differentiated caching strategies for different types of index nodes. It identifies and prioritizes caching of hot-spot nodes (nodes with high access frequency) at upper layers of the index tree, while allowing lower-layer leaf nodes to be accessed from storage when needed. This local quality approach optimizes cache usage by storing only the most frequently accessed index nodes locally, reducing cache management complexity while maintaining improved data access performance for the majority of operations

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP4722939A1Method and apparatus for determining remote index node, device, and medium
Publication Date: 2026.04.08 ZTE CORP
  • EP4722939A1 patent drawingFigure 1~2
  • EP4722939A1 patent drawingFigure 3
  • EP4722939A1 patent drawingFigure 4~5

AI summary

The present application discloses a method, apparatus, device for determining a remote index node, and a medium, belonging to the technical field of distributed storage. A range index cache is constructed for a compute node in a distributed architecture, and the cached index node is constructed in the range index cache. A target cached node covering a range of the determination request of the remote index node is determined by searching, and the remote index node in the storage nodes are determined according to the target cached node.