Durable network coded distributed storage

Abstract

In distributed cloud storages fault tolerance is achieved by regenerating the lost data from the surviving clouds. Recent studies suggest using maximum distance separable (MDS) network codes in cloud storage systems to allow efficient and reliable recovery after node faults. MDS codes are designed to use a substantial number of repair nodes and rely on centralized management and a static fully connected network between the nodes. However, in highly dynamic environments, like edge caching in communication networks or peer-to-peer networks, the nodes and the communication links availability is very volatile. In these scenarios MDS codes functionality is limited. In this paper we study a non-MDS network coded approach, which operates in a decentralized manner and requires a small number of repair nodes for node recovery. We investigate long-term behavior of the modeled system and demonstrate, analytically and numerically, the durability gains over uncoded storage.

Staying alive — Network coding for data persistence in volatile networks

“…The helper data for node repairs is drawn from small node groups, but no specific node in a group is essential for the repair. The scheme was proposed by the authors in [5]. Via numerical simulations, we demonstrate the DRNC performance with various model parameters and explain the effects observed.…”

“…In order to find the storage life time, it is enough to analyze the dynamics of the rank of W t . Per Theorem 1 from [5], the expected rank of W t is bounded by…”

On the Effectiveness of Recoding-based Repair in Network Coded Distributed Storage

“…To verify this assumption, we have performed and described further experiments in Subsection 5.3. Recently, Abdrashitov et al [15] investigated the long-term behavior of a distributed storage system using a model very similar to the one used in this paper. One of their main results was to define approximations on data survival based on Markov chains.…”