Failure Detector of Perfect P Class for Synchronous Hierarchical Distributed Systems

Introduction

In distributed systems failure detectors are used to maintain information about the operational states of other processes. Information provided by a failure detector is assumed unreliable because it can suspect a correct process or not suspect a faulty process. The operational status information of a process provided by two failure detectors at different processes may differ (Cortinas, 2011). In such scenarios completeness and accuracy are the two properties to assess the reliability of failure detectors. Completeness has been further defined into two variations: strong and weak; while, accuracy has been defined into four variations: strong, weak, eventual strong, and eventual weak (Chandra & Toueg, 1996). The strong completeness represents that eventually every process that crashes is permanently suspected by every correct process. Whereas, strong accuracy represents that no process is suspected before it crashes. A failure detector that satisfies strong completeness and strong accuracy properties belong to the Perfect P class. Similarly, there are eight pairs, each pair forming a new failure detector class (see Table 1) formed by selecting one of the two completeness properties and one of the four accuracy properties.

Completeness	Accuracy
	Strong	Weak	Eventual Strong	Eventual Weak
Strong	Perfect	Strong	Eventually Perfect	Eventually Strong
Weak		Weak		Eventually Weak