Programming Sagas in SOCK

2008 Sixth European Conference on Web Services, 2008

Dynamic fault handling is a new approach for dealing with fault management in service-oriented applications. Fault handlers, termination handlers and compensation handlers are installed at execution time instead of being statically defined. In this paper we present this programming style and our implementation of dynamic fault handling in JOLIE, providing finally a nontrivial example of its usage.

2008

Service oriented computing (SOC) allows for the composition of services which communicate using uni-directional notification or bidirectional request-response primitives. Most service orchestration languages proposed so far provide also primitives to handle faults and compensations. The interplay between fault handling and request-response invocations is nontrivial since, for instance, faults should be notified to the request-response communication partners in order to compensate also the remote activities. Our work is motivated by the difficulties encountered in programming, using current orchestration languages, some fault handling strategies. We propose as a solution an orchestration programming style in which fault and compensation handlers are dynamically installed. We show the adequacy of our proposal defining its semantics, and proving that it satisfies some expected high-level properties. Finally, we also show how to apply dynamic handler installation in a nontrivial automotive scenario.

Lecture Notes in Computer Science, 2006

Telecommunication systems should have a high degree of availability, i.e., high probability of correct and timely provision of requested services. To achieve this, correctness of software for such systems and system fault tolerance should be ensured. Application of formal methods helps us to gain confidence in building correct software. However, to be used in practice, formal methods should be well integrated into existing development process. In this paper we propose a formal model-driven approach to development of communicating systems. Essentially our approach formalizes and extends Lyra -a top-down service-oriented method for development of communicating systems. Lyra is based on transformation and decomposition of models expressed in UML2. We formalize Lyra in the B Method by proposing a set of formal specification and refinement patterns reflecting the essential models and transformations of the Lyra service specification, decomposition and distribution phases. Moreover, we extend Lyra to integrate reasoning about fault tolerance in the entire development flow.

2007

Service Oriented Computing (SOC) allows for the composition of services which communicate using unidirectional notification or bidirectional request-response primitives. Most of the service orchestration languages proposed so far provide also primitives to handle faults and manage the subsequent compensation activities. The interplay between these two aspects is non trivial since, for instance, faults should be notified to the request-response communication partners in order to compensate also the remote activities. We first present a simple orchestration scenario requiring a precise distributed fault handling strategy. We show that this strategy cannot be programmed using current orchestration languages; then, we propose a new style for orchestration programming able to specify the required fault management strategy. Finally, we show the generality of our approach by analyzing its properties and applying it to a nontrivial scenario.

2008

Conversations in service-oriented computation are frequently long running. In such a setting, traditional ACID properties of transactions cannot be reasonably implemented, and compensation mechanisms seem to provide convenient techniques to, at least, approximate them. In this paper, we investigate the representation and analysis of structured compensating transactions within a process calculus model, by embedding in the Conversation Calculus certain structured compensation programming abstractions inspired by the ones proposed by Butler, Ferreira, and Hoare. We prove the correctness of the embedding after developing a general notion of stateful model for structured compensations and related results, and showing that the embedding induces such a model.

Lecture Notes in Computer Science, 2010

Compensating CSP (cCSP) is an extension to CSP for modeling long-running transactions. It can be used to specify programs of service orchestration written in a programming language like WS-BPEL. So far, only an operational semantics and a trace semantics are given to cCSP. In this paper, we extend cCSP with more operators and define for it a stable failures semantics in order to reason about non-determinism and deadlock. We give some important algebraic laws for the new operators. These laws can be justified and understood from the stable failures semantics. A case study is given to demonstrate the extended cCSP.

The Journal of Logic and Algebraic Programming, 2005

In this paper we present an algebra of actors extended with mechanisms to model crash failures and their detection. We show how this extended algebra of actors can be successfully used to specify distributed software architectures. The main components of a software architecture can be specified following an object-oriented style and then they can be composed using asynchronous message passing or more complex interaction patterns. We illustrate this process by means of a case study: the specification of a software architecture for intelligent agents which supports a fault tolerant anonymous interaction protocol.

Lecture Notes in Computer Science, 2007

COWS (Calculus for Orchestration of Web Services) is a foundational language for Service Oriented Computing that combines in an original way a number of ingredients borrowed from well-known process calculi, e.g. asynchronous communication, polyadic synchronization, pattern matching, protection, delimited receiving and killing activities, while resulting different from any of them. In this paper, we extend COWS with timed orchestration constructs, this way we obtain a language capable of completely formalizing the semantics of WS-BPEL, the 'de facto' standard language for orchestration of web services. We present the semantics of the extended language and illustrate its peculiarities and expressiveness by means of several examples.

Theoretical Computer Science, 2012

Compensating CSP (cCSP) models long-running transactions. It can be used to specify service orchestrations written in programming languages like WS-BPEL. However, the original cCSP does not allow to model internal (non-deterministic) choice, synchronized parallel composition, hiding or recursion. In this paper, we introduce these operators and define for the extended language a failure-divergence (FD) semantics to allow reasoning about non-determinism, deadlock and livelock. Furthermore, we develop a refinement calculus that allows us to compare the level of non-determinism between long running transactions, and transform specifications for design and analysis.

Electronic Notes in Theoretical Computer Science, 2009

Service Oriented Computing (SOC) is a design paradigm that has evolved from earlier paradigms including object-orientation and component-based software engineering. Important features of services are compositionality, context-independence, encapsulation and re-usability. To support the formal design and analysis of SOC applications recently a number of Service Oriented Calculi have been proposed. Most of them are based on process algebras enriched with primitives specific of service orientation such as operators for manipulating semi-structured data, mechanisms for describing safe client-service interactions, constructors for composing possibly unreliable services and techniques for services query and discovery. In this paper we show a versatile technique for the definition of Structural Operational Semantics of MarCaSPiS, a Markovian extension of one of such calculi, namely the Calculus of Sessions and Pipelines, CaSPiS. The semantics deals in an elegant way with a stochastic version of two-party synchronisation, typical of a service-oriented approach, and with the problem of transition multiplicity while preserving highly desirable mathematical properties such as associativity and commutativity of parallel composition. We also show how the proposed semantics can be naturally used for defining a bisimulation-based behavioural equivalence for MarCaSPiS terms that induces the same equalities as those obtained via Strong Markovian Equivalence.