Storage and repair bandwidth tradeoff for distributed storage systems with clusters and separate nodes

Abstract

In distributed storage systems (DSSs), the optimal tradeoff between node storage and repair bandwidth is an important issue for designing distributed coding strategies to ensure large scale data reliability. The capacity of DSSs is obtained as a function of node storage and repair bandwidth parameters, characterizing the tradeoff. There are lots of works on DSSs with clusters (racks) where the repair bandwidths from intra-cluster and cross-cluster are differentiated. However, separate nodes are also prevalent in the realistic DSSs, but the works on DSSs with clusters and separate nodes (CSN-DSSs) are insufficient. In this paper, we formulate the capacity of CSN-DSSs with one separate node for the first time where the bandwidth to repair a separate node is of cross-cluster. Consequently, the optimal tradeoff between node storage and repair bandwidth are derived and compared with cluster DSSs. A regenerating code instance is constructed based on the tradeoff. Furthermore, the influence of adding a separate node is analyzed and formulated theoretically. We prove that when each cluster contains R nodes and any k nodes suffice to recover the original file (MDS property), adding an extra separate node will keep the capacity if R|k, and reduce the capacity otherwise.

Storage and repair bandwidth tradeoff for heterogeneous cluster distributed storage systems

“…In [16,20], Prakash et al analyzed the tradeoff bounds in detail when the relay nodes were used both in the repair and data collection procedures and connected to different amounts of intra-cluster nodes. On the other hand, the authors of [17,21,22] considered the cluster DSS model without relay nodes, where the intra-cluster and cross-cluster repair bandwidth were analyzed as two parameters in the repair procedure. In [21,23], the authors considered the two-rack model and analyzed the storage and repair bandwidth tradeoff.…”

“…In [22], we investigated the cluster DSS model when the number of help nodes varied, which is different with the constraints in [17,24]. In heterogeneous storage systems, such as peer-to-peer (P2P) networks, sensor networks, hybrid cloud storage systems and content delivery networks [25][26][27], separate nodes are also prevalent.…”

“…In heterogeneous storage systems, such as peer-to-peer (P2P) networks, sensor networks, hybrid cloud storage systems and content delivery networks [25][26][27], separate nodes are also prevalent. For example, in large scale data centers consisting of many racks (clusters) and separate servers, the storage system can be modeled with the cluster DSS model with multiple separate nodes as proposed in [22,28]. In [28], we theoretically analyzed the tradeoff bounds of the cluster DSS model after one separate node was added.…”

“…For example, in large scale data centers consisting of many racks (clusters) and separate servers, the storage system can be modeled with the cluster DSS model with multiple separate nodes as proposed in [22,28]. In [28], we theoretically analyzed the tradeoff bounds of the cluster DSS model after one separate node was added. However, the tradeoff bound for the cluster DSS model with multiple separate nodes has not been characterized and proved theoretically, which is a main motivation of this paper.…”

“…Main definitions used to analyze the HC-DSS model are introduced in Subsection 2.3. Unlike analyzing the cluster DSS model with one separate node [28], we need to investigate all the repair sequences corresponding to different amounts and locations of separate nodes in the HC-DSS model. The main challenge is to find an efficient comparison procedure considering the relationship among multiple separate nodes, which is the main contribution of this paper.…”