23.3. Consistency of data replicas

Data that is needed to support the smooth operation of a system is often replicated. An example is the naming data held by name servers that was discussed in Section 16.11. Replication of data ensures that if a single computer that holds such data fails, another copy exists. Replication may also be used to ensure that there is a copy reasonably near to all points of the system, which is particularly relevant to large-scale systems. Data is therefore replicated for reasons of availability and performance. Applications and services in general may maintain data replicas for these reasons. Communications delay and distributed updates mean that replicas can get out of step. We may have sacrificed consistency to achieve ...

Get Operating Systems: Concurrent and Distributed Software Design now with the O’Reilly learning platform.

O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.