The ISIS Project
Abstract
The ISIS project is now focusing on making the technology of fault- tolerance easier to use. We believe that by adopting what is essentially a 'RISC' approach to software fault-tolerance, it will be possible to address a broader collection of distributed computing problems than we have in the past, making our work useful to practitioners whose applications cannot be addressed efficiently using our current approach (resilient objects). Our plan is to develop a new system that will continue to provide resilient objects at a high level, but will also include support for fault-tolerant process groups at a lower level. This lower level will be directly accessible to programmers, and much of our own software will reside within it, including a collection of fault- tolerant services embodying specialized distributed algorithms, such as a shared memory mechanism. Also included is a paper entitled: Communication Support for Reliable Distributed Computing, which describes a collection of communication primitives integrated with a mechanism for handling process failure and recovery. These primitives facilitate the implementation of fault-tolerant process groups, which can be used to provide distributed services in an environment subject to non-malicious crash failures.
Document Details
- Document Type
- Technical Report
- Publication Date
- May 04, 1986
- Accession Number
- ADA167756
Entities
People
- Ken Birman
Organizations
- Cornell University