Deadlock avoidance in shared-memory multithreaded programs is receiving increased attention as multicore architectures and parallel programming are becoming more prevalent. In our on-going project, called Gadara, the objective is to control the execution of multithreaded programs in order to avoid deadlocks by using techniques from discrete-event control theory. In this project, Petri nets are employed to model parallel programs. This paper formally defines the class of Petri nets that emerges from modeling multithreaded programs, called Gadara nets. Gadara nets are related to, but different from, other classes of nets that have been characterized in deadlock analysis of manufacturing systems. The contributions of this paper include: (i) formal definition of Gadara nets and of controlled Gadara nets; (ii) a behavioral analysis of Gadara nets for liveness and reversibility using siphons; and (iii) identification of a convexity-type property for the set of live markings.

Abstract-We investigate the synthesis of optimal livenessenforcing control policies for Gadara nets, a special class of Petri nets that arises in the modeling of the execution of multithreaded computer programs for the purpose of deadlock avoidance. We consider maximal permissiveness as the notion of optimality. Deadlock-freeness of a multithreaded program corresponds to liveness of its Gadara net model. We present a new control synthesis algorithm for liveness enforcement of Gadara nets that need not be ordinary. The algorithm employs structural analysis of the net and synthesizes monitor places to prevent the formation of a special class of siphons, termed resource-induced deadlymarked siphons. The algorithm also accounts for uncontrollable transitions in the net in a minimally restrictive manner. The algorithm is generally an iterative process and converges in a finite number of iterations. It exploits a covering of the unsafe states that is updated at each iteration. The proposed algorithm is shown to be correct and maximally permissive with respect to the goal of liveness enforcement.

Abstract-Computer hardware is moving from uniprocessor to multicore architectures. One problem arising in this evolution is that only parallel software can exploit the full performance potential of multicore architectures, and parallel software is far harder to write than conventional serial software. One important class of failures arising in parallel software is circularwait deadlock in multithreaded programs. In our on-going Gadara project, we use a special class of Petri nets, called Gadara nets, to systematically model multithreaded programs with lock allocation and release operations. In this paper, we propose an efficient optimal control synthesis methodology for ordinary Gadara nets that exploits the structural properties of Gadara nets via siphon analysis. Optimality in this context refers to the elimination of deadlocks in the program with minimally restrictive control logic. We formally establish a set of important properties of the proposed control synthesis methodology, and show that our algorithms never synthesize redundant control logic. We conduct experiments to evaluate the efficiency and scalability of the proposed methodology, and discuss the application of our results to real-world concurrent software.

No abstract

Service composition has received significant attention in the research community, and the focus has been on service semantics and composition algorithms. Surprisingly, the problem of representation of the composition outcome has been largely ignored. Ad-hoc workflows are often employed, which typically sacrifice alternative paths and parallelism for the sake of simple representation. In this paper, we show how theory of regions, which was originally developed to derive Petri nets from finite state automata, can be applied to find the optimal representation of composition. To apply the theory, we first propose an automaton-based composition framework that incorporates most existing composition techniques without changing the service semantics or its description language. Then based on the special requirements of the composition representation, we develop our own Petri net synthesis algorithm that combines the benefits of two well known algorithms from the theory of regions. We demonstrate that AND/OR workflow nets can limit the concurrency even for simple input/output based service composition, while our Petri net representation is optimal in terms of flexibility and parallelism. Our experimental evaluations include a case study on composing Google Checkout Service, and the study on Oracle BPEL samples, for which our algorithm obtains better concurrent representations for almost all non-trivial cases.