Abstract
This chapter introduces schemes for replication and concurrency control that were originally developed in the context of distributed file systems but which could become prevalent in synchronous groupware, as well.
A central focus of attention is voting schemes such as majority consensus (with and without primary site), weighted voting, write-all-read-any (also known as read-one-write-all, ROWA for short), voting with witnesses (including volatile witnesses, and with and without leading minority), available-copy, dynamic voting (with update sites cardinality and linear ordering, respectively), voting-class, multidimensional voting, and hierarchical voting schemes, especially the hierarchical quorum consensus and tree quorum. In addition we discuss the coding scheme and the grid protocol.
For each scheme we analyze its availability.
This is a preview of subscription content, log in via an institution to check access.
Preview
Unable to display preview. Download preview PDF.
References
In the original proposal of the grid protocol, the permutation of the column indices is not fully random. Rather, the creation of the permutation considers which columns contain nodes that have been unavailable during the attempt to lock the C-cover. However, for simplicity we shall not consider this strategy here. The interested reader will find a discussion of the more complex strategy in the paper by Cheung et al. (1992).
Author information
Authors and Affiliations
Rights and permissions
Copyright information
© 2000 Springer-Verlag Berlin Heidelberg
About this chapter
Cite this chapter
Borghoff, U.M., Schlichter, J.H. (2000). Replication and Concurrency Control. In: Computer-Supported Cooperative Work. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-662-04232-8_5
Download citation
DOI: https://doi.org/10.1007/978-3-662-04232-8_5
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-08631-1
Online ISBN: 978-3-662-04232-8
eBook Packages: Springer Book Archive