Re: Distributed Join

In article <802275189snz@copse.demon.co.uk> Eoin Woods <Eoin@copse.demon.co.uk> writes: >From: Eoin Woods <Eoin@copse.demon.co.uk> >Subject: Re: Distributed Join >Date: 4 Jun 1995 16:10:52 +0100 >In article <MAL.95Jun3100013@yavanna> mal@winternet.com "Michael Leo" writes: >> You could just buy Ingres/Star and use normal everyday SQL, not knowing >> which table is on which node. You can do this from ANY tool, not just >> call-level or embedded C. You can do this from ANY third party tool >> that works against plain old Ingres. Ingres/Star deals with the >> "distriby" stuff behind the Ingres front-end. >> >> Granted the product is not perfect, but its been able to do this for >> the last 5 years, at least. >> >> One has to ask, "Why can't Sybase?" >Well actually, Sybase *can* ! The product to do this is OMNI and it supports >distributed joins across most data sources (including, I think, some >non relational ones). >(Yes, I am a Sybase employee - I'm just posting from home). >Eoin. >-- >Eoin Woods, Watford, Hertfordshire, WD1 1PD, UK >Email: Eoin@copse.demon.co.uk Both oracle and ingres have done a much better job of making it possible to have multi-database joins that are transparent. Oracle has been doing it for about as long as ingres. However anyone considering this for production systems should be aware that nobody is optimizing distributed joins yet. Unfortunately this means that there is no value in making it transparent , since it wouldn't perform worth a shit.