Abstract
In modern computer systems, distributed systems play an increasingly important role, and modeling and verification are crucial in their development. The specificity of many systems requires taking this into account in real time, as time dependencies significantly affect the system’s behavior, when achieving the goals of its processes or with adverse phenomena such as deadlocks. The natural features of distributed systems include the asynchrony of actions and communication, the autonomy of nodes, and the locality of behavior, i.e., independence from any global or non-local features. Most modeling formalisms are derived from parallel centralized systems, in which the behavior of components depends on the global state or the simultaneous achievement of certain states by components. This approach is unrealistic for distributed systems. This article presents the formalism of a timed integrated model of distributed systems that supports all of the mentioned features. The formalism is based on the relation between the states of the distributed nodes and the messages of distributed computations, called agents. This relation creates system actions. A specification in this formalism can be translated into timed automata, the most popular formalism for specifying and verifying timed parallel systems. The translation rules ensure that the semantics of T-IMDS and timed automata are consistent, allowing use of the Uppaal validator for system verification. The development of general formulas for checking the deadlock freedom and termination efficiency allows for automated verification, without learning temporal logics and time-dependent formulas. An important and rare feature is the finding of partial deadlocks, because in a distributed system a common situation occurs in which some nodes/processes are deadlocked, while others work. Examples of checking timed distributed systems are included.
Highlights
The formalism of the distributed systems specification was developed at the Institute of Computer Science, Warsaw University of Technology
Asynchrony is very much needed between cooperating distributed nodes, because synchronous behavior requires some knowledge of the global state, which is difficult to obtain in a distributed system
It should be noted that IMDS is not an automata-based model, it can be represented as a collection of automata, in various ways; the paper [40] formally presents the conversion of IMDS to node automata or to agent automata
Summary
Identifying partial (sometimes called local) deadlocks and checking the distributed termination with model checking techniques [6] requires the designer to have some knowledge of temporal logic, since features should be expressed in terms of the elements of the particular system under verification [7,8]. The reason for this is that the deadlock is often identified as a global state, with no outgoing transitions [9].
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.