Re: Informix vs. Sybase vs. Oracle vs. (gasp) MS SQL Server

[ 1 - I've snipped, what I believe, were a whole bunch of unrelated newsgroups from this note. I've left only the comp.database.* 2 - If you disagree with what I've written, let's be constructive. ] In article <3351f3b1.8299223@gate.idg.no>, Nils.Myklebust@idg.no writes: > ichudov@algebra.com (Igor Chudov @ home) wrote: > > :I have been asked to do a comparison of captioned database engines, > :to choose a database server for a medium size database (less than 2GB, > :about 60-100 tables). > I can do a *moderate* comparison between Informix and Sybase (32 bit port): Sybase 11.0.X Informix 7.22 ------------- ------------- o ~2 gig max for data cache o 1 gig max for data cache Drawback Informix. With a smaller data cache there's less data you can cache in memory. o variable I/O read/writes: o based on platform, only 2/4K read/writes - 2/4/8/16K although on the log there are group writes Drawback Informix. I/O is *always* the bottleneck. Allowing 2/4K only is very limiting. Oracle does it even better... allowing up to 128K per I/O. Excellent for DSS and/or table scans. o private log cache manipulation o internal group commits, no management allowed by the dba Drawback Informix. By allowing the DBA to micro-manage the RDBMS you can tune for high performance. The log is always the bottleneck and allowing more "knobs" is excellent. o table partitioning is limited o heavy support for table fragmentation Drawback Sybase. Informix has excellent "logical" striping. You can fragment (stripe) your table based on key values or round-robin. Very nice! o command line interface: o no command line interface. You isql/sqsh have to use their gui. Ick! Minor drawback Informix. Personally, I hate having to use their GUI. It's cumbersome because it's ascii based so it's not even a GUI. o virtual architecture is based o the RDBMS is comprised of many processes on multiple engines. One engine per CPU. Granularity, IMHO, is not fine enough. Drawback Sybase. I really like Informix's idea of having many smaller lightweight processes to do the work of the RDBMS. This really makes sense as machines with many processors are coming alive. It's easier to distribute the workload across multiple CPU's. o device striping not supported o a DBSPACE may be comprised of striped chunks Drawback Sybase. This is a really neat feature of Informix to allow the DBA to stripe a DBSPACE across chunks (data devices). Overall: ======== Due to the limitation of the data cache (max of 1 gig) and the lack of variable I/O read/writes I'd pick Sybase over Informix as of this writing (4.14.97). For a heavy DSS application I'd probably pick Oracle but I don't know enough about Oracle to really say but with a current max of 128K per I/O.... well... As for stability, I believe that you'll find both products adequate. You'll find a site that says that they've had no problems and another site saying that they have had nothing but problems... Hope this helps. -- Pablo Sanchez | wk: 415.933.3812| pg: 800.930.5635 -or- pablo_p@pager.sgi.com --------------+-----------------+-------------------------------------------- pablo@sgi.com ... when mailing me, place "not spam" in the Subject