QoS of Timeout-based self-Tuned Failure Detectors in iSCSI

Unreliable failure detectors have been an important abstraction to build dependable communication applications over iSCSI storage systems subject to faults. Their implementations are commonly based on timeouts to ensure iSCSI processes termination. However, for systems built on the Internet, it is hard to estimate this time value because traffic variations. In order to increase the quality of service (QoS), self- tuned failure detectors dynamically adapt their timeouts to the communication delay behavior added of a safety margin. In this paper, we evaluate the QoS of a failure detector for predictor and safety margins in iSCSI storage system. As the results show, to improve the QoS, we must consider the relation between the pair predictor-margin. Furthermore, performance and accuracy requirements should be considered for a suitable relationship.