Erasure code backup system for data security

Abstract

With the development of distributed storage, it is critical to determine how to place the copies of data in different locations with the consideration of network performance and data security. The Named Data Networking (NDN) proposes a solution to cache data everywhere along the routing path toward data consumers, so we do not need to manually place the data copies. However, its multi-copy cache strategy wastes storage space and does not consider remote backup to defense physical hazards. It tends to place backup copies closer to each others, so some copies does not have chance to be accessed and physical hazards can destroy all copies in close geographic location. To address the issue, this paper proposes a Advanced Remote Backup Data System (ARBDS ), which brings erasure coding technology and remote backup concept to distributed storage. The erasure coding can achieve the same reliability at a smaller storage cost with less data redundancy, while remote backup concept protects data from physical hazards. With the consideration on reliability and transmission performance, we design an algorithm named erasure code backup algorithm (ECBA) to find the backup nodes with the 57% shorter average length of routing paths to the consumer compared to existing solutions, provide the 92% higher available bandwidth and keep backup copies far from each other to prevent physical hazards.