Hernán C. Melgratti

Multiparty sessions in SOC

Service oriented applications feature interactions among several participants over the network. Mechanisms such as correlation sets and two-party sessions have been proposed in the literature to separate messages sent to different instances of the same service. This paper presents a process calculus featuring dynamically evolving multiparty sessions to model interactions that spread over several participants. The calculus also provides primitives for service definition/invocation and for structured communication in order to highlight the interactions among the different concepts. Several examples from the SOC area show the suitability of our approach.

A connector algebra for P/T nets interactions

A quite flourishing research thread in the recent literature on component-based system is concerned with the algebraic properties of various kinds of connectors for defining well-engineered systems. In a recent paper, an algebra of stateless connectors was presented that consists of five kinds of basic connectors, plus their duals. The connectors can be composed in series or in parallel and employing a simple 1-state buffer they can model the coordination language Reo. Pawel Sobocinski employed essentially the same stateful extension of connector algebra to provide semantics-preserving mutual encoding with some sort of elementary Petri nets with boundaries. In this paper we show how the tile model can be used to extend Sobocinskis approach to deal with P/T nets, thus paving the way towards more expressive connector models.

Connector algebras, petri nets, and BIP

In the area of component-based software architectures, the term connector has been coined to denote an entity (e.g. the communication network, middleware or infrastructure) that regulates the interaction of independent components. Hence, a rigorous mathematical foundation for connectors is crucial for the study of coordinated systems. In recent years, many different mathematical frameworks have been proposed to specify, design, analyse, compare, prototype and implement connectors rigorously. In this paper, we overview the main features of three notable frameworks and discuss their similarities, differences, mutual embedding and possible enhancements. First, we show that Sobocinskis nets with boundaries are as expressive as Sifakis et al.s BI(P), the BIP component framework without priorities. Second, we provide a basic algebra of connectors for BI(P) by exploiting Montanari et al.s tile model and a recent correspondence result with nets with boundaries. Finally, we exploit the tile model as a unifying framework to compare BI(P) with other models of connectors and to propose suitable enhancements of BI(P).

Resolving Non-determinism in Choreographies

Resolving non-deterministic choices of choreographies is a crucial task. We introduce a novel notion of realisability for choreographies ---called whole-spectrum implementation--- that rules out deterministic implementations of roles that, no matter which context they are placed in, will never follow one of the branches of a non-deterministic choice. We show that, under some conditions, it is decidable whether an implementation is whole-spectrum. As a case study, we analyse the POP protocol under the lens of whole-spectrum implementation.

Behaviour, Interaction and Dynamics

The growth and diffusion of reconfigurable and adaptive systems motivate the foundational study of models of software connectors that can evolve dynamically, as opposed to the better understood notion of static connectors. In this paper we investigate the interplay of behaviour, interaction and dynamics in the context of the BIP component framework, here denoted BI(P), as we disregard priorities. We introduce two extensions of BIP: 1) reconfigurable BI(P) allows to reconfigure the set of admissible interactions, while preserving the set of interacting components; 2) dynamic BI(P) allows to spawn new components and interactions during execution. Our main technical results show that reconfigurable BI(P) is as expressive as BI(P), while dynamic BI(P) allows to deal with infinite state systems. Still, we show that reachability remains decidable for dynamic BI(P).

A Formal Analysis of the Global Sequence Protocol

The Global Sequence Protocol (GSP) is an operational model for replicated data stores, in which updates propagate asynchronously. We introduce the GSP-calculus as a formal model for GSP. We give a formal account for its proposed implementation, which addresses communication failures and compact representation of data, and use simulation to prove that the implementation is correct. Then, we use the GSP-calculus to reason about execution histories and prove ordering guarantees, such as read my writes, monotonic reads, causality and consistent prefix. We also prove that GSP extended with synchronous updates provides strong consistency guarantees.

On correlation sets and correlation exceptions in ActiveBPEL

Correlation sets are a programming primitive that allows instance identification in orchestration languages. A correlation set is a set of properties (i.e., values carried on by messages) that are used to associate each received message with a process instance: every time a service receives a message, it explores its content and determines a service instance that should handle the received message. Based on a concrete implementation, this paper proposes a formal model for correlation sets accounting for correlation exceptions. We also investigate different type systems aimed at ensuring that orchestrators are free from some kind of correlation exceptions.

Abstract Processes in Orchestration Languages

Orchestrators are descriptions at implementation level and may contain sensitive information that should be kept private. Consequently, orchestration languages come equipped with a notion of abstract processes , which enable the interaction among parties while hiding private information. An interesting question is whether an abstract process accurately describes the behavior of a concrete process so to ensure that some particular property is preserved when composing services. In this paper we focus on compliance, i.e, the correct interaction of two orchestrators and we introduce two definitions of abstraction: one in terms of traces, called trace-based abstraction , and the other as a generalization of symbolic bisimulation, called simulation-based abstraction . We show that simulation-based abstraction is strictly more refined than trace-based abstraction and that simulation-based abstraction behaves well with respect to compliance.

Concurrency and Probability: Removing Confusion, Compositionally

Assigning a satisfactory truly concurrent semantics to Petri nets with confusion and distributed decisions is a long standing problem, especially if one wants to resolve decisions by drawing from some probability distribution. Here we propose a general solution based on a recursive, static decomposition of (occurrence) nets in loci of decision, called structural branching cells (s-cells). Each s-cell exposes a set of alternatives, called transactions. Our solution transforms a given Petri net into another net whose transitions are the transactions of the s-cells and whose places are those of the original net, with some auxiliary structure for bookkeeping. The resulting net is confusion-free, and thus conflicting alternatives can be equipped with probabilistic choices, while nonintersecting alternatives are purely concurrent and their probability distributions are independent. The validity of the construction is witnessed by a tight correspondence with the recursively stopped configurations of Abbes and Benveniste. Some advantages of our approach are that: i) s-cells are defined statically and locally in a compositional way; ii) our resulting nets faithfully account for concurrency.

A Denotational View of Replicated Data Types

“Weak consistency” refers to a family of properties concerning the state of a distributed system. One of the key issues in their description is the way in which systems are specified. In this regard, a major advance is represented by the introduction of Replicated Data Types (rdts), in which the meaning of operators is given in terms of two relations, namely, visibility and arbitration. Concretely, a data type operation is defined as a function that maps visibility and arbitration into a return value. In this paper we recast such standard approaches into a denotational framework in which a data type is seen as a function that maps visibility into admissible arbitrations. This characterisation provides a more abstract view of RDTs that (i) highlights some of the implicit assumptions shared in operational approaches to specification; (ii) accommodates underspecification and refinement; (iii) enables a categorical presentation of RDT and the development of composition operators for specifications.