Path: newssvr20.news.prodigy.com!newsmst01.news.prodigy.com!prodigy.com!logbridge.uoregon.edu!feed2.news.rcn.net!rcn!nntp.abs.net!uunet!dca.uu.net!snewsf0.syd.ops.aspac.uu.net!news1.optus.net.au!optus!spool01.syd.optusnet.com.au!spool.optusnet.com.au!not-for-mail From: "Noons" Newsgroups: comp.databases.oracle.server References: <3228903a.0309040520.72522c3@posting.google.com> <3f577fce$1@rutgers.edu> <3f589016$0$14563$afc38c87@news.optusnet.com.au> <3f59eca8$0$273$cc9e4d1f@news.dial.pipex.com> <3f5a8f12$2$14558$afc38c87@news.optusnet.com.au> <3F5CA2E2.D3EAAC83@remove_spam.peasland.com> Subject: Re: Oracle performance with Microsoft Project Date: Tue, 9 Sep 2003 21:26:09 +1000 Organization: nunovyabizness Pty Ltd X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 6.00.2600.0000 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 Lines: 57 Message-ID: <3f5db8d4$0$14559$afc38c87@news.optusnet.com.au> NNTP-Posting-Host: 211.29.173.224 X-Trace: 1063106772 14559 211.29.173.224 Xref: newssvr20.news.prodigy.com comp.databases.oracle.server:242682 "Brian Peasland" wrote in message news:3F5CA2E2.D3EAAC83@remove_spam.peasland.com... > > Not that I'm a huge SQL Server fan.....but you can do many of the same > things in SQL Server to tune the engine that you can do in Oracle. Oh no you cannot. Not the "same things". Nowhere near. > You > can tune your engine by tuning the I/O, i.e. physical placement of > datafiles. Well, that's not much the same thing as Oracle where I can even tune the size of the data block. (not that it's much use nowadays, but...). Besides, the tablespace concept and the datafile location in SS is not quite the same as in Oracle. I'd hate to have to I/O tune a SS database with 5000 tables (assuming you can have that many tables + their indexes without the darn thing carking) without access to some very serious SAN hardware... Let's not even mention running two major apps in the same database instance (or its SS equivalent). There just aren't any facilities to adequately partition the database. > The engine can be tuned with statistics that the optimizer > uses in creating an execution plan of the SQL statement. You mean the optimizer can be influenced with stats. That's not exactly the same as the "engine". What about when the stats are not enough and *still* produce wrong plans? Not much there other than hacks. I'd kill for a profiling facility like Oracle's. All they have really is the "Oracle mode", which is a rough equiv of RBO. > And you can > even tune the engine by modifying initialization parameters. All half a dozen of them... > Granted, > with Oracle you can do much, much more. But SQL Server does allow some > of this as well. Nowhere near enough to be of any use. And M$ will actively discourage you from fiddling with parameters anyway. And managing a fully logged and archived log database is not an easy task at the best of times, particularly when you really need to recover. You better have enough disk to store all the needed archived logs on-line or else. You DO remember which they were and in what order, don't you?... ;) No, the thing is w-a-a-a-a-y behind what you can do in Oracle. Not even within comparison. However, this is not to say it won't get better. It will. It's MOL where Oracle was in V5 in terms of tuning. -- Cheers Nuno Souto wizofoz2k@yahoo.com.au.nospam