Distributed DBMS Reliability
We have referred to “reliability” and “availability” of the database a number of times so far without defining these terms precisely. Specifically, we mentioned these terms in conjunction with data replication, because the principle method of building a reliable system is to provide redundancy in system components. We also claimed in Chapter 1 that the distribution of data enhances system reliability. However, the distribution of the database or the replication of data items is not sufficient to make the distributed DBMS reliable. A number of protocols need to be implemented within the DBMS to exploit this distribution and replication in order to make operations more reliable.
KeywordsBuffer Manager Network Partitioning State Transition Diagram Site Failure Mean Time Between Failure
Unable to display preview. Download preview PDF.