next up previous
Next: Mirroring Up: The Design of Previous: Selection of Workstations

Reliability

In an distributed system, a workstation may crash at any time. In the case of a remote memory server crash, we would like to be able to complete the execution of the application, and recover its lost pages. To provide this level of reliability, some form of redundancy must be used. We explore three different policies: mirroring, parity, and parity caching.





Evangelos Markatos
Fri Mar 24 14:41:51 EET 1995