A Novel Node Selection Method in Wireless Distributed Edge Storage Based on SDN and Multi-attribute Decision Model

TL;DR

This paper introduces a new node selection method for wireless distributed edge storage that considers network and load conditions, leveraging SDN and multi-attribute decision models to enhance performance and scalability.

Contribution

It proposes an innovative SDN-based edge storage architecture with a multi-attribute decision algorithm for node selection, improving performance and scalability in wireless edge storage systems.

Findings

Significant improvement in high-load write performance.

Enhanced scalability and adaptability of the storage system.

Cost savings in hardware deployment.

Abstract

The distributed edge storage system can store data collected at the edge of the network in a decentralised manner, with low latency, high security, and flexibility. Traditional edge-distributed storage systems only consider one single factor, such as node capacity, when storing data, ignoring network and storage node load conditions that affecting the system's read/write performance. At the same time, it could be more scalable in the widely used wireless terminal application scenarios. To tackle these challenges, this paper proposes an innovative software-defined edge storage architecture based on SDN (Software-Defined Networking) and SMB (Server Message Block) protocols, A data storage node selection algorithm that integrates the network state and storage node load state is designed based on multi-attribute decision model, and a system prototype is realised in conjunction with 5G wireless communication technology. Experimental results demonstrate significant improvements in the performance of high-load write operations compared to traditional edge-distributed storage systems. The proposed wireless distributed edge storage system also demonstrates superior scalability and adaptability, effectively addressing the challenge of limited system scalability and improving compatibility with edge scenarios in mobile applications. In addition, it results in cost savings in hardware deployment and presents a promising advancement in edge storage technology.