Browse by author
Lookup NU author(s): Professor Brian RandellORCiD
An Algorithm is presented for providing continued processing in a distributed system in the presence of network partitioning. It involves a protocol for automatic detection and resolution of mutual inconsistencies between replicated data items in partitioned distributed systems which does not require a prior knowledge of the. semantics of operations on data structures. The algorithm, which is based on the use of majority and minority partitions, allows autonomous processing in individual nodes when a network partitions thereby providing continued processing despite network failures. Actual (as opposed to potential) conflicts between the operations in different partitions are detected and resolved during a merge action. The resolution of actual conflicts avoids the unnecessary UNDO processing which arises when potential rather than just those conflicts which actually occur are resolved. An implementation of the proposed algorithm in a hierarchical distribted system is discussed.
Author(s): Anyanwu JA, Randell B
Publication type: Report
Publication status: Published
Series Title: Computing Laboratory Technical Report Series
Year: 1985
Pages: 22
Report Number: 204
Institution: Computing Laboratory, University of Newcastle upon Tyne
Place Published: Newcastle upon Tyne
URL: http://www.cs.ncl.ac.uk/publications/trs/papers/204.pdf
