A precise definition of interaction behavior between services is a prerequisite for successful business-to-business integration. Service choreographies provide a view on message exchanges and their ordering constraints from a global perspective. Assuming message sending and receiving as one atomic step allows to reduce the modelers’ effort. As downside, problematic race conditions resulting in deadlocks might appear when realizing the choreography using services that exchange messages asynchronously. This paper presents typical issues when desynchronizing service choreographies. Solutions from practice are discussed and a formal approach based on Petri nets is introduced for identifying desynchronizable choreographies.

This chapter describes a design methodology for business processes and workflows that focuses first on “business artifacts”, which represent key (real or conceptual) business entities, including both the business-relevant data about them and their macro-level lifecycles. Individual workflow services (a.k.a. tasks) are then incorporated, by specifying how they operate on the artifacts and fit into their lifecycles. The resulting workflow is specified in a particular artifact-centric workflow model, which is introduced using an extended example. At the logical level this workflow model is largely declarative, in contrast with most traditional workflow models which are procedural and/or graph-based. The chapter includes a discussion of how the declarative, artifact-centric workflow specification can be mapped into an optimized physical realization. 1.

Abstract. A choreography describes the interaction between services. It may be used for specification purposes, for instance serving as a contract in the design of an interorganizational business process. Typically, not all describable interactions make sense which motivates the study of the realizability problem for a given choreography. In this paper, we show that realizability can be traced back to the problem of controllability which askes whether a service has compatible partner processes. This way of thinking makes algorithms for controllability available for reasoning about realizability. In addition, it suggests alternative definitions for realizability. We discuss several proposals for defining realizability which differ in the degree of coverage of the specified interaction. 1

Abstract. We propose a methodology for statically predicting the possible interaction patterns of services within a given choreography. We focus on choreographies exploiting the event notification paradigm to manage service interactions. Control Flow Analysis techniques statically approximate which events can be delivered to match the choreography constraints and how the multicast groups can be optimised to handle event notification within the service choreography. 1

Abstract- Converged services alignment (i.e., orchestration and choreography) has been a dominant topic of several regulatory initiatives. These initiatives were successful to some extent, but had faced also various difficulties, that prevented any final specification grade. Thus, this paper revises the converged services orchestration and choreography concerns that apply for an effective enterprise management and develops a framework that enables service composition considering multiple interconnected perceptions. Furthermore, since the regulatory initiatives in this area have not been built on top of a defined and all-encompassing conceptual foundation, we present a viable alternative to make up for this drawback. The paper also shapes a service provisioning arrangement intended to identify requirements and concepts to be addressed by and integrated into the specification framework.

Modeling of application- and middleware-layer interaction protocols

Modeling and analysis of interactions among services is a crucial issue in Service-Oriented Computing. Composing Web services is a complicated task which requires techniques and tools to verify that the new system will behave correctly. In this paper, we first overview some formal models proposed in the literature to describe services. Second, we give a brief survey of verification techniques that can be used to analyse services and their interaction. Last, we focus on the realizability and conformance of choreographies. inria-00539017, version 1- 23 Nov 2010 1

Abstract. We put forward an interface and component algebra through which we characterise fundamental structures that support service-oriented design independently of the specific formalisms that may be adopted to provide models for languages or analysis tools. We view services as an interface mechanism that can be superposed over a component infrastructure, what is sometimes referred to as a ‘service overlay’. The component algebra consists of networks of processes that interact asynchronously through communication channels. A service interface offers properties to potential clients and requires properties of external services that, at run time, may need to be discovered and bound to the orchestration of the service. We define what it means for an asynchronous relational net to orchestrate a service interface and prove a number of compositionally results that relate the operations of both algebras. One of the major results of the paper is the characterisation of a sub-class of asynchronous relational nets over which we can guarantee that, when binding, through their interfaces, a client and a supplier service, the composition of the orchestrations of the two services is consistent, i.e., both services can work together as interconnected. 1