Mirco Tribastone

SENSORIA Patterns: Augmenting Service Engineering with Formal Analysis, Transformation and Dynamicity

The IST-FET Integrated Project Sensoria is developing a novel comprehensive approach to the engineering of service-oriented software systems where foundational theories, techniques and methods are fully integrated into pragmatic software engineering processes. The techniques and tools of Sensoria encompass the whole software development cycle, from business and architectural design, to quantitative and qualitative analysis of system properties, and to transformation and code generation. The Sensoria approach takes also into account reconfiguration of service-oriented architectures (SOAs) and re-engineering of legacy systems.

Scalable Differential Analysis of Process Algebra Models

The exact performance analysis of large-scale software systems with discrete-state approaches is difficult because of the well-known problem of state-space explosion. This paper considers this problem with regard to the stochastic process algebra PEPA, presenting a deterministic approximation to the underlying Markov chain model based on ordinary differential equations. The accuracy of the approximation is assessed by means of a substantial case study of a distributed multithreaded application.

The PEPA eclipse plugin

The PEPA Eclipse Plug-in supports the creation and analysis of performance models, from small-scale Markov models to large-scale simulation studies and differential equation systems. Whichever form of analysis is used, models are expressed in a single highlevel language for quantitative modelling, Performance Evaluation Process Algebra (PEPA).

Automatic extraction of PEPA performance models from UML activity diagrams annotated with the MARTE profile

Recent trends in software engineering lean towards modelcentric development methodologies, a context in which the UML plays a crucial role. To provide modellers with quantitative insights into their artifacts, the UML benefits from a framework for software performance evaluation provided by MARTE, the UML profile for model-driven development of Real Time and Embedded Systems. MARTE offers a rich semantics which is general enough to allow different quantitative analysis techniques to act as underlying performance engines. In the present paper we explore the use of the stochastic process algebra PEPA as one such engine, providing a procedure to systematically map activity diagrams onto PEPA models. Independent activity flows are translated into sequential automata which co-ordinate at the synchronisation points expressed by fork and join nodes of the activity. The PEPA performance model is interpreted against a Markovian semantics which allows the calculation of performance indices such as throughput and utilisation. We also discuss the implementation of a new software tool powered by the popular Eclipse platform which implements the fully automatic translation from MARTE-annotated UML activity diagrams to PEPA models.

The PEPA Plug-in Project

We present a GUI-based tool supporting the stochastic process algebra PEPA with modules for performance evaluation through Markovian steady-state analysis, fluid flow analysis, and stochastic simulation.

Stochastic Process Algebras

In this paper we report on progress in the use of stochastic process algebras for representing systems which contain many replications of components such as clients, servers and devices. Such systems have traditionally been difficult to analyse even when using high-level models because of the need to represent the vast range of their potential behaviour. Models of concurrent systems with many components very quickly exceed the storage capacity of computing devices even when efficient data structures are used to minimize the cost of representing each state. Here, we show how population-based models that make use of a continuous approximation of the discrete behaviour can be used to efficiently analyse the temporal behaviour of very large systems via their collective dynamics. This approach enables modellers to study problems that cannot be tackled with traditional discrete-state techniques such as continuous-time Markov chains.

Non-functional properties in the model-driven development of service-oriented systems

Systems based on the service-oriented architecture (SOA) principles have become an important cornerstone of the development of enterprise-scale software applications. They are characterized by separating functions into distinct software units, called services, which can be published, requested and dynamically combined in the production of business applications. Service-oriented systems (SOSs) promise high flexibility, improved maintainability, and simple re-use of functionality. Achieving these properties requires an understanding not only of the individual artifacts of the system but also their integration. In this context, non-functional aspects play an important role and should be analyzed and modeled as early as possible in the development cycle. In this paper, we discuss modeling of non-functional aspects of service-oriented systems, and the use of these models for analysis and deployment. Our contribution in this paper is threefold. First, we show how services and service compositions may be modeled in UML by using a profile for SOA (UML4SOA) and how non-functional properties of service-oriented systems can be represented using the non-functional extension of UML4SOA (UML4SOA-NFP) and the MARTE profile. This enables modeling of performance, security and reliable messaging. Second, we discuss formal analysis of models which respect this design, in particular we consider performance estimates and reliability analysis using the stochastically timed process algebra PEPA as the underlying analytical engine. Last but not least, our models are the source for the application of deployment mechanisms which comprise model-to-model and model-to-text transformations implemented in the framework VIATRA. All techniques presented in this work are illustrated by a running example from an eUniversity case study.

Automatic Translation of UML Sequence Diagrams into PEPA Models

The UML profile for modeling and analysis of real time and embedded systems (MARTE) provides a powerful, standardised framework for the specification of non-functional properties of UML models. In this paper we present an automatic procedure to derive PEPA process algebra models from sequence diagrams (SD) to carry out quantitative evaluation. PEPA has recently been enriched with a fluid-flow semantics facilitating the analysis of models of a scale and complexity which would defeat Markovian analysis.

Stochastic process algebras

In this tutorial we give an introduction to stochastic process algebras and their use in performance modelling, with a focus on the PEPA formalism. A brief introduction is given to the motivations for extending classical process algebra with stochastic times and probabilistic choice. We then present an introduction to the modelling capabilities of the formalism and the tools available to support Markovian based analysis. The chapter is illustrated throughout by small examples, demonstrating the use of the formalism and the tools.

Exact fluid lumpability for Markovian process algebra

We study behavioural relations for process algebra with a fluid semantics given in terms of a system of ordinary differential equations (ODEs). We introduce label equivalence, a relation which is shown to induce an exactly lumped fluid model, a potentially smaller ODE system which can be exactly related to the original one. We show that, in general, for two processes that are related in the fluid sense nothing can be said about their relationship from stochastic viewpoint. However, we identify a class of models for which label equivalence implies a correspondence, called semi-isomorphism, between their transition systems that are at the basis of the Markovian interpretation.