Ten Years of Performance Evaluation for Concurrent Systems Using CADP

Abstract

Abstract. This article comprehensively surveys the work accomplished during the past decade on an approach to analyze concurrent systems qualitatively and quantitatively, by combining functional verification and performance evaluation. This approach lays its foundations on semantic models, such as Imc (Interactive Markov Chain) and Ipc (Interactive Probabilistic Chain), at the crossroads of concurrency theory and mathematical statistics. To support the approach, a number of software tools have been devised and integrated within the Cadp (Construction and Analysis of Distributed Processes) toolbox. These tools provide various functionalities, ranging from state space generation (Caesar and Exp.Open), state space minimization (Bcg Min and Determinator), numerical analysis (Bcg Steady and Bcg Transient), to simulation (Cunctator). Several applications of increasing complexity have been successfully handled using these tools, namely the Hubble telescope lifetime prediction, performance comparison of mutual exclusion protocols, the Scsi-2 bus arbitration protocol, the Send/Receive and Barrier primitives of Mpi (Message Passing Interface) implemented on a cache-coherent multiprocessor architecture, and the xSTream multiprocessor data-flow architecture for embedded multimedia streaming applications.

Quantitative Timed Analysis of Interactive Markov Chains

“…Untimed quantitative reachability analysis of IMCs has been handled in [11]; timed reachability in [30]. Other related work is on continuous-time Markov decision processes (CTMDPs).…”

“…Imca has a simple input format that facilitates its usage as a back-end tool for other tools that generate IMCs from high-level model specifications such as AADL, DFTs, Prism reactive modules, and so on. It supports the bcg-format, such that it accepts state spaces generated (and possibly minimized) using the CADP toolbox [11]; CADP supports a LOTOS-variant for the compositional modeling of IMCs and compositional minimization of IMCs.…”

“…CTMCs can sometimes be obtained from IMCs by applying weak bisimulation minimization; however, if this does not suffice, semantic restrictions on the IMC level are imposed to ensure full probabilism. The CADP toolbox [11] supports the compositional generation, minimization, and standard CTMC analysis of IMCs. In this paper, we focus on the quantitative timed analysis of arbitrary IMCs, in particular of those, that are non-deterministic and can be seen as stochastic realtime 1 1 2 -player games.…”

SMT-Based Bisimulation Minimisation of Markov Models

“…"), using Prism we were not able to build the state space let alone perform the actual verification. 4 Here, the crucial advantage of the Smt-based quotienting becomes apparent: since it avoids building the full state space of the original model, it shortens computation times while reducing the required memory.…”

“…The main drawback of bisimulation quotienting algorithms, however, is that they typically require the entire reachable state space. Techniques to alleviate this effect include the use of data structures based on binary decision diagrams (BDDs) to reduce storage costs [21,20,17], and compositional minimisation techniques [4]. This paper takes a radically different approach: we extract the bisimulation quotient directly from a high-level description using Smt (satisfiability modulo theories) solvers.…”