A Data-Centric Approach for Scalable State Machine Replication
Data replication is a key design principle for achieving reliability, high-availability, survivability and load balancing in distributed computing systems. The common denominator of all existing replication systems is the need to keep replicas consistent. The main paradigm for supporting replicated data is active replication, in which replicas execute the same sequence of methods on the object in order to remain consistent. This paradigm led to the definition of State Machine Replication (SMR) [29.8], [29.13]. The necessary building block of SMR is an engine that delivers operations at each site in the same total order without gaps, thus keeping the replica states consistent.
Unable to display preview. Download preview PDF.
- 29.1R. Boichat, P. Dutta, S. Frolund and R. Guerraoui. Deconstructing Paxos. Technical Report DSC ID:200106, Communication Systems Department (DSC), École Polytechnic Fédérale de Lausanne (EPFL), January 2001. Available at http://dscwww.epfl.ch/EN/publications/documents/tr01006.pdf.
- 29.3G. Chockler and D. Malkhi. Active disk Paxos with infinitely many processes. In Proceedings of the 21st ACM Symposium on Principles of Distributed Computing (PODC’02 ), July 2002. To appear.Google Scholar
- 29.4G. Chockler, D. Malkhi and M. K. Reiter. Backoff protocols for distributed mutual exclusion and ordering. In Proceedings of the 21st International Conference on Distributed Computing Systems, pages 11–20, April 2001.Google Scholar
- 29.6E. Gafni and L. Lamport. Disk Paxos. In Proceedings of 14th International Symposium on Distributed Computing (DISC’2000), pages 330–344, October 2000.Google Scholar
- 29.10W. K. Lo and V. Hadzilacos. Using failure detectors to solve consensus in asynchronous shared-memory systems. In Proceedings of the 8th InternationalWorkshop on Distributed Algorithms (WDAG), Springer-Verlag LNCS 857:280–295, Berlin, 1994.Google Scholar
- 29.12J. P. Martin, L. Alvisi and M. Dahlin. Minimal Byzantine Storage. In Proceedings of the 16th International Conference on DIStribued Computing (DISC’02), pages 311–325, October 2002Google Scholar