On the Quality of Service of Failure Detectors

We study the Quality of Service (QoS) of failure detectors. By QoS, we mean a specification that quantifies 1) how fast the failure detector detects actual failures and 2) how well it avoids false detections. We first propose a set of QoS metrics to specify failure detectors for systems with probabilistic behaviors, i.e., for systems where message delays and message losses follow some probability distributions. We then give a new failure detector algorithm and analyze its QoS in terms of the proposed metrics. We show that, among a large class of failure detectors, the new algorithm is optimal with respect to some of the QoS metrics. Given a set of failure detector QoS requirements, we show how to compute the parameters of our algorithm so that it satisfies these requirements and we show how this can be done even if the probabilistic behavior of the system is not known. We then present some simulation results that show that the new failure detector algorithm provides a better QoS than an algorithm that is commonly used in practice. Finally, we suggest some ways to make our failure detector adaptive to changes in the probabilistic behavior of the network.