An Adaptive Failure Detection Method with Controllable QoS for Distributed Storage System

Abstract

According to the situation that different application asks for different requirement of storage node's performance, an adaptive failure detection method with controllable QoS for distributed storage system is proposed. The delay characteristic of heartbeat message is summarized. The relational expressions of neighbor heartbeat message's arrival time in the different network state are established. The QoS parameters formulas of different network state are proposed. The controllability of QoS parameters is analyzed. The algorithm of the method is designed. The experimental results show that the method realizes the balance control between detection speed and detection accuracy by setting safe remaining time with a certain value, and has adaptability that it can have a high detection speed but fewer mistakes.