GlobalComp: Models, semantics, logics and technologies for global computing
A Global Computing (GC) system is a programmable infrastructure distributed at worldwide scale and available globally. GC distinguishes from traditional distributed computing in terms of scalability, connectivity, heterogeneity, and autonomy. In the GC context, important and new issues are given by the ability of dealing with dynamic and unpredictable changes of their network environment (e.g. availability of network connectivity, lack of resources, node failures, network reconfigurations and so on). At the same time, older problems - such as the expressiveness and modularity at the level of programming languages – become more acute, in the sense that the new technologies should reach new standards to allow us to cope with the complexity and diversity that are implied by the emerging model of the GC.
The aim of the researchers that participate the GlobalComp project is to develop prototype implementations and new mathematical models for GC that enable formal verification. We will study systems compatible with the notion of GC. The most prominent example is that of the Web space. In particular, the growing complexity of the entire collection of Internet services brings the need for a change in the view of the Web space as just a database. Thus, it appears the opportunity to harness the available computational power. In the emerging GC vision, the Web should become a programmable space. An important research theme of this project will be the study of GC systems from the perspective of the P2P paradigm.
Our aim is to study, elaborate and introduce formal models (that enable formal reasoning) and technologies for a number of GC key issues: distributed networks with dynamic topologies, data and code mobility, security, coordination and control. We will perform experiments on P2P architectures, and develop prototype implementations for (concepts of) multiparadigm languages or calculi, trust engines, context aware applications, interactive systems, etc. For the theoretical aspects of our research we will use our previous results and experience with various operational, denotational and axiomatic models of concurrency.
The GlobalComp project involves four academic partners – the Technical University of Cluj-Napoca (UTCN), the “Babes-Bolyai” University of Cluj-Napoca (UBB), the University of Bucharest (UB), and the A.I.Cuza University of Iasi (UAIC) - and one partner from the industry - Siemens PSE (SIEMENS).
The research activities of the GlobalComp group at UAIC will be focused mainly on the foundational aspects of GC. We will use our previous experience and results in process algebras (pi-calculus and its extensions: Dpi, tDpi; mobile ambients), membrane systems, bigraphs, virtual machines, topologies, categories and domain theory.
At UB the main focus will be on interactive computing and the use of spatio-temporal logics for reasoning about their behavior. We intend to develop structured programming languages for interactive systems with a sound mathematical and logical foundation, and to use them in the development of GC applications, particularly for Web services.
At UTCN we will develop prototype implementations for various notions that are relevant in the GC context, with main focus on specific programming languages concepts. For the theoretical aspects of our research we will use our previous results and experience with the use of continuations, monads and type systems.
At UBB and SIEMENS we intend to focus on the formal modeling of Peer-to-Peer (P2P) GC systems, and also on an implementation of such a GC system in joint work with the SIEMENS partner. The main proposed goal, is to transfer on the Web such a P2P system.