[PhD] - An Infrastructure for Mobile Service-Oriented Computing Encoded on a Process Calculus
Sep 2002 - Dec 2006
Service-Oriented computing and mobile agents are two of the major current technologies for Internet applications. In the service-oriented paradigm, computation is modeled as components that communicate by providing and requiring services. Mobile agents are computations that have the ability to travel through a network, by halting their execution, saving their state and then restoring it in a new host. This thesis introduces the Mob scripting programming language to develop applications based on mobile agents, with service-oriented resource discovery. It is our belief that combining both paradigms in a programming language provides a useful tool for high-level programming of mobile agents. For instance, using a service-oriented language, programmers are not required to keep track of agent names; applications can be made resilient since the same service may be provided transparently by several agents in the network, which in turn makes it easier to implement load-balancing strategies to improve efficient processing of client requests. Mob combines both paradigms by providing mobile agents and services as the main abstractions of the language. We start by defining the Mob language, providing its syntax and, in a informal manner, its semantics. We then present some applications that illustrate the use of the language constructs and its expressiveness, From an implementation point of view, the compiler is divided in two stages. The first is a front-end that outputs code in DiTyCO (Distributed Typed Concurrent Objects), a language based on a process calculus with primitives for distribution and mobility of resources. This allows us to use both a compiler and a run-time system, we have developed for DiTyCO, respectively, as the back-end of the Mob compiler and as the run-time system for Mob.
