From total order to database replication

Y. Amir1, C. Tutu1
1Department of Computer Science, Johns Hopkins University, Baltimore, MD, USA

Tóm tắt

This paper presents in detail an efficient and provably correct algorithm for database replication over partitionable networks. Our algorithm avoids the need for end-to-end acknowledgments for each action while supporting network partitions and merges and allowing dynamic instantiation of new replicas. One round of end-to-end acknowledgments is required only upon a membership change event such as a network partition. New actions may be introduced to the system at any point, not only while in a primary component. We show how performance can be further improved for applications that allow relaxation of consistency requirements. We provide experimental results that demonstrate the efficiency of our approach.

Từ khóa

#Partitioning algorithms #Transaction databases #Engines #Computer science #Availability #Computer crashes #Throughput #Delay #Distributed computing #Prototypes

Tài liệu tham khảo

10.1145/3149.214121 golding, 1992, Weak-consistency group communication and membership gray, 1993, Transaction Processing: concepts and techniques, Data Management Systems hadzilacos, 1993, Fault-tolerant broadcasts and related problems, Distributed Systems holliday, 2000, Database replication using epidemic update, Technical Report TRCS00-01, 19 10.1145/78922.78926 keidar, 1994, A highly available paradigm for consistent object replication, Institute of Computer Science 10.1145/212433.212468 10.1145/363951.363955 10.1109/DSN.2001.941398 10.1109/TSE.1979.234180 amir, 1998, The spread wide area group communication system, Technical Report CNDS-98-4 stanoi, 1998, Using broadcast primitives in replicated databases, Proceedings of the 18th IEEE International Conference on Distributed Computing Systems '98, 148 amir, 2002, Practical wide-area database replication, Technical Report CNDS-2002-2 10.1145/320128.320131 10.1109/ICDCS.2002.1022299 10.1145/41840.41841 10.1145/41457.37515 amir, 1995, Replication Using Group Communication over a Partitioned Network 10.1145/377769.377776 amir, 1993, A highly available application in the Transis environment, Lecture Notes in Computer Science, 774, 125 10.1109/ICDCS.1994.302392 pedone, 1999, The Database State Machine and Group Communication Issues 10.1007/3-540-40026-5_21 10.1145/119995.115856 pedone, 1998, Exploiting atomic broadcast in replicated databases, Proceedings of Eu-roPar (EuroPar'98) skeen, 1982, A quorum-based commit protocol, Berkeley Workshop Distributed Data Management and Comput Networks 10.1145/98163.98167