home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!cimshop!davidm
- From: davidm@consilium.com (David S. Masterson)
- Newsgroups: comp.databases.theory
- Subject: Re: Optimistic Concurrency Control
- Message-ID: <DAVIDM.92Aug12133625@consilium.com>
- Date: 12 Aug 92 20:36:25 GMT
- References: <713419958snx@cascade.demon.co.uk> <DAVIDM.92Aug10113706@consilium.com>
- <1992Aug11.050133.23420@redbrick.com>
- Sender: root@cimshop.UUCP
- Organization: Consilium Inc., Mountain View, California
- Lines: 22
- In-reply-to: jfr@redbrick.com's message of 11 Aug 92 05:01:33 GMT
- X-Posting-Software: GNUS 3.13 [ NNTP-based News Reader for GNU Emacs ]
-
- >>>>> On 11 Aug 92 05:01:33 GMT, jfr@redbrick.com (Jon Rosen) said:
-
- > The advantages of optimistic concurrency [...] are the high level of
- > concurrency without lock contention that can be achieved. We at Red Brick
- > find this particularly interesting because query servers do not have the
- > need to provide a high level of concurrency control for transaction
- > processing. We are willing to sacrifice TPS rates in exchange for high
- > availability and query performance. Much of the overhead of TPS-bound
- > systems occurs inside the lock manager, particularly if a seperate process
- > is involved in handling locking activity.
-
- Rdb/VMS implements read-only snapshot files (I imagine others do to) that
- allow queries to operate concurrently with write transactions. How does this
- affect your view of the usefulness of optimistic locking?
- --
- ====================================================================
- David Masterson Consilium, Inc.
- (415) 691-6311 640 Clyde Ct.
- davidm@consilium.com Mtn. View, CA 94043
- ====================================================================
- "When the government attempts to regulate everything, all is lost."
- -- Thibaudeau
-