home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky comp.databases:8674 comp.databases.oracle:2630 comp.databases.sybase:561
- Path: sparky!uunet!cs.utexas.edu!swrinde!zaphod.mps.ohio-state.edu!pacific.mps.ohio-state.edu!linac!att!cbnewsc!cbfsb!att-out!rutgers!ucla-cs!ucla-mic!agsm!ddruker
- From: ddruker@agsm.ucla.edu (Daniel Druker)
- Newsgroups: comp.databases,comp.databases.oracle,comp.databases.sybase
- Subject: Re: 500'000 records - who does best?
- Message-ID: <1992Dec23.132155.15738@mic.ucla.edu>
- Date: 23 Dec 92 21:21:54 GMT
- References: <1992Dec21.182948.16832@bernina.ethz.ch>
- Organization: Anderson Graduate School of Management, UCLA
- Lines: 40
- Nntp-Posting-Host: risc.agsm.ucla.edu
-
- In article <1992Dec21.182948.16832@bernina.ethz.ch> Florian.Gutzwiller@open.ch writes:
- >If you'd have to realize a server client based database with Suns as servers
- >and NeXT as clients. Single records would be an average of 30KB in size and
- >there would be a number of 20 concurrent users.
- >
- >Would you choose Sybase or Oracle ?
- >Who has experience with similar solutions/environments ?
-
- I'd choose Oracle, hands down, based on backup and recovery issues. 500k recs
- of 30kb each means you have a pretty large database. Figure in additional
- space for indexes and sorting space and you might be talking 30 Gigabytes
- or more. Sybase will have serious problems backing up a database of this
- size. It will probably take DAYS to perform a backup, load, or restore.
- Oracle treats large databases much better, and includes lots of parallel
- stuff which really helps on multi-processors. I expect query speeds will
- be similar since you're really only gated by how fast the disk spins,
- and with a database of this size you probably wont want to allow many
- (or ANY) full table scans, ad hoc stuff, etc..., especially on a sun
- with a limited number of procs...
-
- Oracle has a very large database group which they can put you in touch with
- to get some practical experience of users with I think up to 100 or so gigs
- of data.
-
- Hope this helps,
-
- - Dan
-
- Daniel Druker
- Anderson Graduate School of Management at UCLA
-
-
- ----------------------------------------------------------------------------
- | Dan Druker |
- | agsm mail : ddruker |
- | internet : ddruker@agsm.ucla.edu |
- | oracle*mail : unix:ddruker@agsm.ucla.edu |
- ----------------------------------------------------------------------------
-
- Disclaimer: None. I'm a student now and I don't care what you think.
-