Estimating Event Lifetimes for Distributed Runtime Verification

Runtime system verification has been proposed as a form of dynamic verification of software systems which can be applied in settings where complete static verification or exhaustive system testing is not practical. Runtime verification checks properties against runtime events generated during the operation of a system. Current approaches to runtime verification assume that runtime events are time-stamped by a single clock and, thus, can be totally ordered. They also assume that events are received by the reasoning engine in the same order as they have been produced. These assumptions are apparently true only in systems with a single clock. In this paper, we present the extension of a framework for runtime verification which can monitor distributed systems, in which events are produced by different components, each having its own clock. 1.