Pietro Cenciarelli

From Sequential to Multi-Threaded Java: An Event-Based Operational Semantics

A structural operational semantics of a non trivial sublanguage of Java is presented. This language includes dynamic creation of objects, blocks, and synchronization of threads. First we introduce a simple operational description of the sequential part of the language, where the memory is treated as an algebra with suitably axiomatized operations. Then, the interaction between threads via a shared memory is described in terms of structures, called “event spaces,” whose wellformedness conditions formalize directly the rules given in the Java language specification. Event spaces are included in the operational judgements to develop the semantics of the full multi-threaded sublanguage, which is shown to extend the one for sequential Java conservatively. The result allows sequential programs to be reasoned about in a simplified computational framework without loss of generality.

Ambient Graph Rewriting

We investigate synchronized hyperedge replacement (SHR) as general framework for distributed programming and system design. We propose a slender version of SHR which dramaticaly reduces the mathematical overhead of the original proposal G. Ferrari, U. Montanari, and E. Tuosto. A LTS semantics of ambients via graph synchronization with mobility. ITCS, 2001] and use it to interpret the distributed CCS J. Riely and M. Hennessy. Distributed Processes and Location Failures. Theoretical Computer Science, 266:693--735, 2001] and the calculus of Mobile Ambients L. Cardelli and A.D. Gordon. Mobile Ambients. Theoretical Computer Science, 1(240):177--213, 2000] in a uniform semantic framework. The encodings are bisimulations. A tool for supporting distributed system design and analysis is presented. The tool, which adopts the slender SHR as intermediate language, integrates model checking techniques within the framework of (distributed) program development.

Depletable channels: dynamics and behaviour

A simple model of multi-hop communication in ad-hoc networks is considered. Similar models are often adopted for studying energy efficiency and load balancing of different routing protocols. We address an orthogonal question never considered by the networking community: whether, regardless of specific protocols, two networks may be considered as equivalent from the viewpoint of the communication service they provide. In particular, we consider equivalent two networks with identical maximum and minimum inhibiting flow, and prove that this notion of equivalence coincides with a standard trace-based notion of equivalence borrowed from the theory of concurrency. We finally study the computational complexity of the proposed equivalence and discuss possible alternatives.

An Algebraic View of Program Composition

We propose a general categorical setting for modeling program composition in which the call-by-value and call-by-name disciplines fit as special cases. Other notions of composition arising in denotational semantics are captured in the same framework: our leading examples are nondeterministic call-by-need programs and nonstrict functions with side effects. Composition of such functions is treated in our framework with the same degree of abstraction that Moggis categorical approach based on monads allows in the treatment of call-by-value programs. By virtue of such abstraction, interesting program equivalences can be validated axiomatically in mathematical models obtained by means of modular constructions.

Inefficiencies in network models: A graph-theoretic perspective

Abstract We consider three network models where information items flow from a source to a sink node: flow networks, depletable channels, and traffic networks. We start with the standard model of flow networks; we characterise graph topologies that admit non-maximum saturating flows, under some capacity-to-edge assignment. We then consider a model where routing is constrained by energy available on nodes in finite supply (like in Smartdust) and efficiency is related to energy consumption and again to maximality of saturating flows. Finally, we consider a traffic model for selfish routing, where efficiency is related to latency at a Wardrop equilibrium. We show that all these forms of inefficiency yield different classes of graphs (apart from in the acyclic case, where the first and the last forms generate the same class). Interestingly, in all cases inefficient graphs can be made efficient by removing edges; this resembles a well-known phenomenon, called Braesss paradox.

Network Applications of Graph Bisimulation

Synchronising Graphsis a system of parallel graph transformation designed for modeling process interaction in a network environment. We propose a theory of context-freesynchronising graphs and a novel notion of bisimulation equivalence which is shown to be a congruence with respect to graph composition and node restriction. We use this notion of equivalence to study some sample network applications, and show that our bisimulation equivalence captures notions like functional equivalence of logical switches, equivalence of channel implementations and level of fault tolerance of a network.

Rational Unification in 28 Characters

We present a case study where Synchronising Graphs, a system of parallel graph transformation, are used to solve the syntactic unification problem for first order rational terms (with possibly infinite unifier). The solution we offer is efficient, that is quasi-linear, and simple: a program of 28 characters.

Verifying a compiler optimization for Multi-Threaded Java

The specification for the object-oriented concurrent language Java [3] is rather loose with respect to the interaction of shared memory and the local working memories of different threads. This allows maximal freedom in the language implementation. Such freedom is reflected in the semantics provided in [2], where threads-memory interaction is formalized in terms of structures called event spaces. Two kinds of memories are described in the Java specification: a “normal” memory and a more liberal one, where values can sometimes be stored even before they are produced as results of computation. Here we compare two structural operational semantics of a sublanguage of Java modelling the two types of memory. The two semantics share the same set of operational rules but put different requirements (expressed as first order theories) on the notion of event space. We prove a result which is informally stated in

A Polynomial-Time Algorithm for Detecting the Possibility of Braess Paradox in Directed Graphs

A directed multigraph is said vulnerable if it can generate Braess paradox in traffic networks. In this paper, we give a graph–theoretic characterisation of vulnerable directed multigraphs. Analogous results appeared in the literature only for undirected multigraphs and for a specific family of directed multigraphs. The proof of our characterisation provides the first polynomial time algorithm that checks if a general directed multigraph is vulnerable in