An Effective Data Recovery Approach for Distributed Storage based on Interference Alignment
Jingyao Zhang · 2022
Node breakdown is very common in distributed storage systems, which will result in data loss or damage. Regenerating codes can reduce the network bandwidth cost by broken data recovery. Interference Alignment (IA) code is a type of regenerating code that can minimize the required network bandwidth for data regeneration. The original IA code could only recover a single breakdown node. In this work, we shall propose a solution based on the IA code that can jointly repair multiple breakdown nodes. We will derive the Repair Matrix that can regenerate multiple lost data blocks through simple multiplication. Based on this scheme, a multi-failure recovery algorithm based on IA code will be put forward. Moreover, the network bandwidth required for data recovery under different scenarios will be studied through both theoretical and numerical analysis.