The Storage vs Repair Bandwidth Trade-off for Multiple Failures in Clustered Storage Networks

TL;DR

This paper investigates the balance between storage efficiency and repair bandwidth in clustered storage systems, especially when repairing multiple node failures, providing bounds and insights into optimal repair strategies.

Contribution

It characterizes the optimal storage-bandwidth trade-off for multiple failures in clustered storage, including exact and functional repair, revealing key conditions affecting system capacity.

Findings

Trade-off same as single failure when t divides (m-ell)

Exact repair at MBR can have less file size than functional repair

More local helpers do not always increase capacity under functional repair

Abstract

We study the trade-off between storage overhead and inter-cluster repair bandwidth in clustered storage systems, while recovering from multiple node failures within a cluster. A cluster is a collection of $m$ nodes, and there are $n$ clusters. For data collection, we download the entire content from any $k$ clusters. For repair of $t \geq 2$ nodes within a cluster, we take help from $\ell$ local nodes, as well as $d$ helper clusters. We characterize the optimal trade-off under functional repair, and also under exact repair for the minimum storage and minimum inter-cluster bandwidth (MBR) operating points. Our bounds show the following interesting facts: $1)$ When $t|(m-\ell)$ the trade-off is the same as that under $t=1$, and thus there is no advantage in jointly repairing multiple nodes, $2)$ When $t \nmid (m-\ell)$, the optimal file-size at the MBR point under exact repair can be strictly less than that under functional repair. $3)$ Unlike the case of $t=1$, increasing the number of local helper nodes does not necessarily increase the system capacity under functional repair.