A method for trade-off analysis and explicit construction of universal rack-aware regenerating codes

By optimizing the matrix structure and encoding strategy of rack-aware regeneration codes in distributed storage systems, the problem of insufficient cross-rack repair bandwidth optimization under rack failures in existing technologies is solved. The minimum storage and minimum cross-rack repair bandwidth under single-node and single-rack failures are achieved, thereby improving the system's fault tolerance performance.

CN114490171BActive Publication Date: 2025-09-26CLOUD LINK TECH (GUANGDONG) CO LTD
2 Cites 0 Cited by

Patent Information

Application Number
CN202210169882.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-02-23
Publication Date
2025-09-26
Estimated Expiration
2042-02-23

AI Technical Summary

Technical Problem

In existing distributed storage systems, when a rack fails, cross-rack repair bandwidth optimization mainly focuses on single-node failures, failing to effectively optimize cross-rack repair bandwidth under rack failures, resulting in a failure to achieve the optimal trade-off between storage and repair bandwidth.

Method used

A trade-off analysis and explicit construction method for a universal rack-aware regenerating code is proposed. By defining data symbol grouping and matrix structure, the cross-rack repair bandwidth under single-node failure and single-rack failure is optimized. The coding strategy of relay nodes and auxiliary racks is adopted to achieve the optimal trade-off between minimum storage and minimum cross-rack repair bandwidth.

Benefits of technology

In the event of a single-node failure or a single-rack failure, the cross-rack repair bandwidth is significantly reduced, providing an optimized solution for minimum storage and minimum cross-rack repair bandwidth, thereby improving the fault tolerance and efficiency of the distributed storage system.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

This invention is applicable to the fields of distributed storage and network coding technology improvements. It provides a method for trade-off analysis and explicit construction of a universal rack-aware regeneration code. In data centers, storage nodes are organized in racks, and cross-rack communication bandwidth is typically much lower than intra-rack communication bandwidth. Two common failures in data centers are single-node failures and single-rack failures. After achieving the minimum cross-rack repair bandwidth for single-node failures, the optimal trade-off between storage and cross-rack repair bandwidth for single-rack failures is described. A universal storage code family, the Generalized Rack-Aware Regeneration Code (GRRC), achieves this optimal trade-off. Two extreme points of GRRC are obtained: the Minimum Storage Generalized Rack-Aware Regeneration Point (MSGRR) and the Minimum Bandwidth Generalized Rack-Aware Regeneration Point (MBGRR).
Need to check novelty before this filing date? Find Prior Art