Decentralized Runtime Verification for LTL Properties Using Global Clock

TL;DR

This paper introduces a decentralized algorithm for runtime verification of LTL properties in distributed systems with a shared global clock, improving efficiency and fault tolerance over centralized methods.

Contribution

It presents a novel sound and complete decentralized algorithm for LTL runtime verification in asynchronous distributed systems with a global clock.

Findings

The algorithm is sound and complete for the specified system model.

Decentralized verification reduces message traffic compared to centralized approaches.

The approach enhances fault tolerance and scalability in distributed runtime verification.

Abstract

Runtime verification is the process of verifying critical behavioral properties in big complex systems, where formal verification is not possible due to state space explosion. There have been several attempts to design efficient algorithms for runtime verification. Most of these algorithms have a formally defined correctness property as a reference and check whether the system consistently meets the demands of the property or it fails to satisfy the property at some point in runtime. LTL is a commonly used language for defining these kinds of properties and is also the language of focus in this paper. One of the main target systems for runtime verification are distributed systems, where the system consists of a number of processes connecting to each other using asynchronous message passing. There are two approaches for runtime verification in distributed systems. The first one consists of centralized algorithms, where all processes send their events to a specific decision-making process, which keeps track of all the events to evaluate the specified property. The second approach consists of distributed algorithms, where processes check the specified property collaboratively. Centralized algorithms are simple, but usually involve sending a large number of messages to the decision-making process. They also suffer from the problem of single point of failure, as well as high traffic loads towards one process. Distributed algorithms, on the other hand, are usually more complicated, but once implemented, offer more efficiency. In this paper, we focus on a class of asynchronous distributed systems, where each process can change its own local state at any arbitrary time and completely independent of others, while all processes share a global clock. We propose a sound and complete algorithm for decentralized runtime verification of LTL properties in these systems.