Two-pass multi-version read consistency

作者: Gary Hallmark , Roger Bamford

DOI:

关键词:

摘要: The present invention describes a method and apparatus that forces all parts of an update transaction to commit automatically using two-phase commit. uses the messages negotiate system number (SCN) is then recorded on each nodes as logical point in time distributed committed. scheme consists "prepare" phase "commit" phase. database environment involves multiple participants (databases). In prepare phase, one N selected "commit point." includes log into which information written states has remaining N-1 enter "prepare phase." lock their databases state they are ready roll back or transaction. mid-point at (nth participant) writes its log. second informed release locks have been set.

参考文章(7)
Ashok K. Agrawala, Mukesh Singhal, A Concurrency Control Algorithm and Its Performance for Replicated Database Systems. international conference on distributed computing systems. pp. 140- 147 ,(1986)
David B. Lomet, Kenneth Wilner, James Johnson, Philip A. Bernstein, System and method for consistent timestamping in distributed computer databases ,(1991)
Nandit Soparkar, Henry F. Korth, Abraham Silberschatz, None, Failure-resilient transaction management in multidatabase IEEE Computer. ,vol. 24, pp. 28- 36 ,(1991) , 10.1109/2.116886
Raymond A. Lorie, Chandrasekaran Mohan, Mir H. Pirahesh, Multiple version database concurrency control system ,(1991)
Miho C, Yojiro C, Koichi C, Nobuhiro C, A distributed database system ,(1991)