Sub-space approach to distributed system development on large heterogeneous distributed environments

  • Authors:
  • Zhili Sun

  • Affiliations:
  • University of Lancaster, Lancaster, U.K.

  • Venue:
  • EW 3 Proceedings of the 3rd workshop on ACM SIGOPS European workshop: Autonomy or interdependence in distributed systems?
  • Year:
  • 1988

Quantified Score

Hi-index 0.00

Visualization

Abstract

It becomes very difficult to implement network transparency on large heterogeneous distributed environments. Each computer system can fail independently because of system autonomy, and a large amount of system status should be maintained and information exchanged to keep the distributed system consistent. The system management can be complicated and performance can suffer.In practice, groups of computers are naturally formed by clustering of according to the sub-groups of an organisation, kinds of computer systems or network structures. Therefore the large heterogeneous distributed system environment is naturally divided into sub-environments, called subspaces. Cooperation between the subspaces and network transparency within the subspaces will be the primary issues. This paper introduce a new approach to cope with those issues. A distributed operating system (DOS) model is designed based on such an approach and the prototype model is implemented by using remote procedure calls (RPCs), such as the SUN RPC[1] and Courier RPC[2] distributed programming facilities and functions on the department's experimental distributed environment. It is argued that the sub-space approach is a promising way to develop a distributed system in a large heterogeneous distributed system environment.