Guixiang Lv
Department of Electrical and Computer Engineering, University of Massachusetts, Dartmouth, MA, USA.

Liudong Xing
Department of Electrical and Computer Engineering, University of Massachusetts, Dartmouth, MA, USA.

Honggang Wang
Department of Electrical and Computer Engineering, University of Massachusetts, Dartmouth, MA, USA.

Hong Liu
Department of Electrical and Computer Engineering, University of Massachusetts, Dartmouth, MA, USA.

DOI https://doi.org/10.33889/IJMEMS.2023.8.1.001

  ;

Abstract

Storage area networks (SANs) are one of the prevalent reliable data storage solutions. However, cascading failures triggered by data overloading have become a major threat to SANs, preventing the desired quality of service from being delivered to users. Based on our preliminary works on studying the impacts of data loading on the reliability performance of SANs, this paper advances the state of the art by implementing node degree-based load redistribution strategies to enhance the SAN reliability, thus mitigating or even preventing the occurrence of cascading failures during the mission time. Load-based and reliability-based node selection rules are considered, which choose nodes with the highest load level and the lowest reliability for load redistribution, respectively. The relationship between data loading and reliability of an individual SAN component is modeled using the accelerated failure-time model with the power law. The SAN reliability is assessed using a combinatorial decision diagram-based approach. The application and effectiveness of the proposed load redistribution strategies are demonstrated and compared through a case study of an SAN with the mesh topology.

Keywords- Storage area network, Cascading failure, Load redistribution, Node degree, Accelerated failure-time model.

Citation

Lv, G., Xing, L., Wang, H., & Liu, H. (2023). Load Redistribution-based Reliability Enhancement for Storage Area Networks. International Journal of Mathematical, Engineering and Management Sciences, 8(1), 1-14. https://doi.org/10.33889/IJMEMS.2023.8.1.001.