Scientific Foundations

Middleware for Autonomous distRibuted applications with context manaGEment

The broad domain of MARGE research is the design and evaluation of models, algorithms, paradigms and tools for building distributed systems and applications. Our historical background is distributed systems. We have worked on object-, component-, agent-, and service-based middleware architectures aimed at supporting the execution of distributed applications in environments of any scale, involving devices of any kind.

In the last years, our work has focused on algorithms and middleware for distributed applications in mobile and pervasive environments. Pervasive applications are hosted by various and heterogeneous devices (including mobile terminals, wireless sensor networks and electronic tags) and various wireless networks (Zigbee, Bluetooth, Wifi). Since their execution environment are variable, pervasive applications have to adapt themselves dynamically to various situations. We propose new paradigms to design, deploy, execute and monitor pervasive applications.

We apply our research to various application domains. These applications allow us to validate new propositions with real problems. Domain applications we have worked on so far are crisis management, mobile e-commerce, ambiant assistant living, multi-players games on mobile devices and mobile learning. We mainly work on collaborative projects with academic and industrial partners (French and European research programs).

You will find below some subjects on which we have worked during the recent years.

  • Middleware and services for consistency management of multiple copies in the context of mobile users in wireless environments.
  • Distributed algorithms for disconnection and partition detectors in mobile and ad~hoc networks.
  • Models and services to choose the most suited assembly of services (or components) in mobile and pervasive environments.
  • Service session continuity to enable distributed applications to migrate some of their components during their execution.
  • Domain specific meta-models for modeling context-aware applications and reconfigure at runtime application context-awareness.
  • Context managers to identify adaptation situations in autonomous computing.
  • Middleware for multi-players games on mobile phones.
  • Mobile agents as a new paradigm for the autonomous deployment of software on various networks (e.g., network of sensors, P2P networks, grid networks).
Back to Computer Science Department Access Map Terms of Service