home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.databases
- Path: sparky!uunet!spool.mu.edu!sdd.hp.com!scd.hp.com!cupnews0.cup.hp.com!dhepner
- From: dhepner@cup.hp.com (Dan Hepner)
- Subject: Re: Hot Standby DBMS's
- Sender: news@cupnews0.cup.hp.com
- Message-ID: <BuDIEM.J80@cup.hp.com>
- Date: Thu, 10 Sep 1992 17:39:58 GMT
- References: <9209091721.AA27050@ludwig.sharebase.com>
- Organization: Hewlett-Packard
- X-Newsreader: Tin 1.1scd1 PL4
- Lines: 26
-
- From: glenn@ludwig.sharebase.com (Glenn Linderman)
-
- >-> Any ideas on how to convince them?
-
- >The point of passing the journal data from the master RM to a Hot Standby RM
- >is performance:
- [but..]
- >As a result, then, you would have an extremely complicated translation
- >system that would probably be as slow as or slower than the original
- >queries. So passing the original queries to both machines in parallel
- >via a TM sounds like a pretty good alternative, and you can have that
- >today.
- >GLenn
-
- What you can't have today is any sensible way to resynchronize
- when an out-of-date copy is to be brought back on-line. We have
- discussed before why the input stream cannot be logged for
- future replay, but the journal stream could be.
-
- I think the first-order performance evaluation is that journal
- data is more consumptive of communication bandwidth, the input
- stream is more consumptive of cycles at the standby machine.
- For most people, a small advantage to the input stream, but
- not enough to drive an entire industry to make that decision.
-
- Dan Hepner
-