GODC: Group-Oriented Distributed Computing

From Navigators

Jump to: navigation, search


  • Sponsor: ESPRIT BRA (and USA NSF), since 1993
  • Coordinator: INESC
  • Partners: INESC, Cornell University
  • Start Date: Nov. 1993
  • Duration: 24 months
  • Keywords: groups, distributed systems, large scale, CSCW
  • Team at INESC: 6 researchers, including Paulo Verissimo

Abstract

Increasing use of distributed systems, with the corresponding decentralization of activities, stimulate the need for structuring those activities around diverse groups of interested participants. The building blocks for group activity are currently the subject of great interest. The notion pervades all layers of a distributed architecture, from multicasting communication infrastructures and group communication protocols, to group management services. Operating system and networking support should evolve to accommodate this reality.

Aims

Both the Navigators team at INESC and the ISIS group at Cornell have a track record of producing group-oriented technologies. They propose to cooperate in order to exchange ideas about their respective work, and possibly open the way for reuse of technologies each one develops. Understanding the requirements of the user applications on one side, and the constraints of the environment on the other, are fundamental pre-conditions to establish: the most adequate group activity building blocks; the functional and correctness dependencies among them; and their efficient implementation in the operating system, from network to applications support environment.

Paradigms and algorithms to assist the solution of these distributed problems have been presented in the recent years. A systematics of group orientation in distributed systems is yet to be developed. It is hoped to achieve results in this area. In particular, about its adaptability to large scale distributed computing systems.

Approach and Methods

This is a strategic research action, where the teams will merge and discuss their experiences in g-o technology. As far as INESC is concerned, this will work both internally to group, from the several ongoing projects (NavTech, NavCim, CoDiCom), and externally, with the BROADCAST (cf. project synopsis elsewhere) partners and Cornell. It is intended to investigate around the 'group' paradigm, in order to discover the adequate properties and structure of each of the building blocks of a g-o system, to respond to the new challenges in distributed computing.

Progress and Results

The cooperation has focused on: (i) models and frameworks for large-scale group support; (ii) a few key technologies - remote and reliable client-server access; group membership and addressing; lightweight groups. In the course of the cooperation, several visits have taken place. Several papers and reports on models and frameworks have been produced. Both have influenced the production of several papers on particular technologies.

Potential

It is expected that research in this area will unveil new techniques to support construction of systems and applications, based on the group paradigm.

Publications

  • Luís Rodrigues, António Casimiro, Paulo Verissimo, “Priority-based totally ordered multicast”, in May 1995. Proceedings of the 3rd IFAC/IFIP workshop on Algorithms and Architectures for Real-Time Control (AARTC'95). Ostend, Belgium, May 1995.

BibTeX

Navigators - GODC project
Personal tools
Navigators toolbox