Abstract
Formal verification of distributed systems is essential, especially in mission-critical systems that cannot be restarted. Such are space systems in which satellites read sensor values and autonomously make actuator decisions based on them, and ground services only set general patterns of behavior. The verification formalism should correspond to the essential characteristics of a distributed system, such as node autonomy and asynchrony of actions and communication, as in our Integrated Model of Distributed Systems (IMDS). It is also crucial that the formalism allows for finding partial deadlocks and checking partial termination, where only a subset of the system nodes is involved while the rest can perform their own tasks at the same time. This article presents the idea of using monitoring agents—observers prepared in the IMDS formalism. Observers check the state of individual system components by polling, allowing verification without knowing the global state of the system. Such an agent is an ideal prototype of a runtime observer that checks if the actual operation of the system corresponds to a design that has previously been proven correct.
Subject
Electrical and Electronic Engineering,Biochemistry,Instrumentation,Atomic and Molecular Physics, and Optics,Analytical Chemistry
Reference36 articles.
1. What Really Heppened on Mars Rover Pathfinder;Jones;Risk Dig.,1997
2. The structure of the “the”-multiprogramming system
3. Model Checking;Clarke,1999
4. Principles of Model Checking;Baier,2008
5. Fairness in Temporal Verification of Distributed Systems
Cited by
3 articles.
订阅此论文施引文献
订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献