Comparative analysis of quality of service and memory usage for adaptive failure detectors in healthcare systems

Failure detection (FD) is an important issue for supporting dependability in distributed healthcare systems to guarantee continuous, safe, secure, and dependable operation, and often is an important performance bottleneck in the event of node failure. FD can be used to manage the health status of communication for delivering telemedicine services, and then to help distributed healthcare system reduce fatal accident rate and increase the reliability and safety of systems. Ensuring acceptable quality of service (QoS) is made difficult by the relative unpredictability of the network environment. In this paper, first, we compare QoS metrics of several adaptive FDs, discuss their properties and their relation, and then propose one optimization over the existing methods, called tuning adaptive margin failure detector (TAM FD), which significantly improves QoS, especially in the aggressive range and when the network is unstable. Second, we address the problem of most adaptive schemes, namely their need for a large window of samples. So we also analyze the impact of memory size on the performance of FDs, and then prove that the presented scheme is designed to use a fixed and very limited amount of memory for the distributed system. Our experimental results over several kinds of networks (Cluster, WiFi, LAN, Intercontinental WAN) show that the properties of the existing adaptive failure detectors, and demonstrate that the optimization is reasonable and acceptable. Furthermore, the extensive experimental results show what is the effect of memory size on the overall QoS of each adaptive failure detector. For our TAM FD, the effect of window size on their QoS is very small and can be negligible.