The Conversation Calculus: A Model of Service-Oriented Computation

Abstract

We present a process-calculus model for expressing and analyzing service-based systems. Our approach addresses central features of the serviceoriented computational model such as distribution, process delegation, communication and context sensitiveness, and loose coupling. Distinguishing aspects of our model are the notion of conversation context, the adoption of a context sensitive, message-passing-based communication, and of a simple yet expressive mechanism for handling exceptional behavior. We instantiate our model by extending a fragment of the π-calculus, illustrate its expressiveness by means of many examples, and study its basic behavioral theory; in particular, we establish that bisimilarity is a congruence. IntroductionWeb services have emerged mainly as a toolkit of technological and methodological solutions for building open-ended collaborative software systems on the Internet. Many concepts that are frequently put forward as distinctive of service-oriented computing, namely, object-oriented distributed programming, long duration transactions and compensations, separation of workflow from service instances, late binding and discovery of functionalities, are certainly not new, at least when considered in isolation. What is certainly new about services is that they are contributing to physically realize (on the Internet) a global, interaction-based, loosely-coupled, model of computation. We would like to better understand in what sense service orientation is to be seen as a new paradigm to build and reason about distributed systems.The main contributions of this work are the development of a process calculus for service-oriented computing based on a novel notion of conversation context, and the study of its basic behavioral theory. In particular, we establish that bisimilarity is a congruence, thus asserting the proper status of the proposed constructions as operators at the level of the behavioral semantics; we believe that such a result has not yet been provided for other related service calculi. Our starting point is an attempt to isolate and clarify essential characteristics of the service-oriented model, in order to propose a motivation from "first principles" of a reduced set of general abstractions for expressing and analyzing service-based systems. We then instantiate our model by modularly extending the static fragment of the π-calculus with conversation contexts, messagepassing communication primitives, and an exception handling mechanism. Some Key Aspects of Service-Oriented ComputingWe identify as key aspects of the service-oriented computational model: distribution, process delegation, communication and context sensitiveness, and loose coupling.Distribution. The purpose of a service relationship is to allow the incorporation of certain activities in a given system, without having to engage local resources and capabilities to support or implement such activities. By delegating activities to an external service provider, which will perform them using its own remote resources and capabilities,...

SLMC: A Tool for Model Checking Concurrent Systems against Dynamical Spatial Logic Specifications

“…The development continued under the support of project IP Sensoria [2], where the tool was included in the Sensoria tool suite, and extended for the verification of service-oriented systems, as described in [17]. In particular, we have also concluded recently further extensions to the tool, namely an extension to the applied π-calculus (for security), and another for checking choreography conformance of service-oriented applications, based on an encoding of the Conversation Calculus [16]. Further information about the Spatial Logic Model Checker may be found in http://ctp.di.fct.unl.pt/SLMC/.…”

Inference of Global Progress Properties for Dynamically Interleaved Multiparty Sessions

“…[6] proposes a sophisticated proof system which builds a wellfounded ordering on events (similar to the approach pursued in [21]) to enforce progress for processes of the Conversation Calculus [20], also in presence of dynamic join and leave of participants. Their progress is guaranteed under the assumption that all communications are matched with sufficient joiners.…”

Conversation Types

“…In this section, we present the syntax and operational semantics of the core Conversation Calculus (CC) [19]. The core CC extends the π-calculus [16] static fragment with the conversation construct n [P ], and replaces channel based communication with context-sensitive message based communication.…”

“…(a) represent communications with the environment, and this represents a conversation identity access; these correspond to the basic actions a process may perform in the context of a given conversation. To capture the observational semantics of processes [19], transition labels register not only the action but also the conversation where the action takes place. So, a transition label λ containing c σ is said to be located at conversation c (or just located), otherwise is said to be unlocated.…”

“…In previous work [19] we introduced the Conversation Calculus (CC), a π-calculus based model for service-oriented computing that builds on the concepts of process delegation, loose-coupling, and, crucially, conversation contexts.…”