Reliability of a Multicomputer Distributed Operating System
Abstract
The project on Reliability of a Multicomputer Distributed Operating System is a three-year research effort in reliability and availability of multicomputer systems with decentralized control. The main objective of this project is to allow a loosely coupled network of computers to cooperate in order to provide a general-purpose non-stop computing facility. The main issues involved with increasing the reliability are replication and multiple (concurrent) activities. This report covers the third research year: August 1, 1987 to July 31, 1988. During this period we investigated four main areas for increasing the reliability. First, we developed a set of algorithms for finding the global average load of a large multicomputer system. The second part is a study of a holographic filing system for a multicomputer with a large number of disks. The third area is a performance study of remote vs. local system calls in the Multicomputer Operating System distributed system. The forth area is an effort to develop mechanisms for supporting massive parallel tasks in a distributed environment. Keywords: Israel, Message passing.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 05, 1988
- Accession Number
- ADA201906
Entities
People
- Amnon Barak
Organizations
- Hebrew University of Jerusalem