Research Report - Distributed Systems
This research addresses two major problems of distributed system design:
1. Interoperability of heterogeneous distributed systems
2. Enhancement of distributed systems to achieve pervasive systems which allow ‘anytime, anywhere’ computing.
Distributed systems are very often large, complex and built on heterogeneous computing platforms (e.g. CORBA, DCOM) which makes cooperation between applications from different platforms, or migration of applications between different platforms, very difficult. Open Distributed Computing (ODP) introduced a framework for building large, autonomous, heterogeneous distributed systems. The group is investigating how this framework can be extended to support interoperability of heterogeneous distributed applications in large enterprises. The main focus of our research is on (i) the creation of a type model able to describe distributed objects and their interfaces in heterogeneous systems in order to support interoperability, (ii) policy based management systems for large, evolving enterprises which allow dynamic introduction of roles and policies and provide dynamic conflict detection and resolution for policies, and (iii) run-time changes to object semantics to allow evolution in distributed applications.
The main thrust of this research is, however, on pervasive systems. The recent advances in wireless technology, the proliferation of mobile devices and advances in embedded systems have created a requirement for ‘anytime, anywhere’ computing in distributed systems. This trend is leading towards the scenario, termed pervasive computing, in which cheap, interconnected computing devices are ubiquitous and capable of supporting users in a wide range of tasks. Pervasive computing environments will differ from current computing environments in many respects, but particularly in scale and complexity. As a result, the success of pervasive computing technologies will rely on a radical design shift. It will require computing applications to become more adaptive, in order to cope with highly dynamic environments and changing user requirements, as well as placing fewer demands on user attention. Accordingly, applications will need to become increasingly sensitive to context (i.e. the circumstances or situation in which a computing task takes place).
One of the great challenges of pervasive computing is to capture the requirements and current status of the computing environment to allow evaluation of changes in this environment and to allow decision-making on how to adapt the whole system to these changes. Such adaptations can be carried out at the various layers of the system (e.g. by adapting user applications or applications' communication streams, adapting the distributed computing middleware or the behaviour of the underlying operating system or communication protocols). Our research focuses on many aspects of building pervasive systems including: context models, adaptation methods, service discovery in pervasive systems, and programming pervasive systems.
Associated Staff
A/Prof Jaga Indulska
Prof Simon Kaplan
Dr Ralf Muhlberger
Prof Maria Orlowska
Dr Peter Sutton
Dr Stephen Viller
A/Prof Xiaofang Zhou
Mr Jihan Zhu
<%@ include file="/include/footer.inc.jsp" %>