2017/08/17 by Vitaly Abdrashitov, Abdrashitov, Vitaly, N. Prakash +3
Computer Science · #Advanced Data Storage Technologies #Distributed systems and fault tolerance #FOS: Computer and information sciences #Information Theory (cs.IT) #Interconnection Networks and Systems
paper · pdf · doi:10.48550/arxiv.1708.05474
openalex publication_date 2017/08/17 · openalex created_date 2019/07/23 · openalex updated_date 2026/07/28
We study the trade-off between storage overhead and inter-cluster repair\nbandwidth in clustered storage systems, while recovering from multiple node\nfailures within a cluster. A cluster is a collection of m nodes, and there\nare n clusters. For data collection, we download the entire content from any\nk clusters. For repair of t \≥ 2 nodes within a cluster, we take help\nfrom \ℓ local nodes, as well as d helper clusters. We characterize the\noptimal trade-off under functional repair, and also under exact repair for the\nminimum storage and minimum inter-cluster bandwidth (MBR) operating points. Our\nbounds show the following interesting facts: 1) When t|(m-\ℓ) the\ntrade-off is the same as that under t=1, and thus there is no advantage in\njointly repairing multiple nodes, 2) When t nmid (m-\ℓ), the optimal\nfile-size at the MBR point under exact repair can be strictly less than that\nunder functional repair. 3) Unlike the case of t=1, increasing the number\nof local helper nodes does not necessarily increase the system capacity under\nfunctional repair.\n