Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: Microsoft claim to fame : The TeraServer on The Web

Re: Microsoft claim to fame : The TeraServer on The Web

From: Scott Cote <scottcote_at_contractor.net>
Date: 1998/07/08
Message-ID: <35A437C3.DEEBC199@contractor.net>#1/1

Ok wetfart - you have made your - ahem point quite grotesqly (spelling).

You might want to add to your diatribe of limitations:

  sql splatter 7.0 doesn't run on anything else other than NT (a big limitation in my book).

SCott

Billy Verreynne wrote:

> I for one has been saying not to use SQL-Server for large databases - and do
> not recommend to run databases larger than 10 to 20GB on SQL-Server.
>
> Now Microsoft has released SQL-Server 7.0 Enterprise - and to show that it
> can handle large databases they've created a 1.something terabyte database
> consisting of 170 million rows of satelite images. Great hey?
>
> http://www.terraserver.microsoft.com/
>
> Well, not really.
>
> A terabyte db with 170 million rows!? That's nothing. We have that many rows
> in a single 30GB table in an Oracle data warehouse. Size of the database
> DOES NOT give ANY indication as to how good performance of that database
> is. When it gets to size, the performance is dependant on hardware issues
> like the speed of disk i/o channels, search and read speed of the disk
> controller, etc.
>
> What's that? SQL-Server users read ahead threads? So fucking what. Whether
> you're accessing 170 million rows that's 1 terabyte in size, or 170 million
> rows that 30GB in size, read ahead techniques do not make a difference when
> it comes to the size of the row. That's dependant on hardware and memory.
>
> I get a response of less than 2 or 3 seconds pulling a single entity's rows
> from our 170 million row table.
>
> The real issues are:
> - how well does SQL-Server handle tables with a billion
> plus rows (real terabyte data warehousing)
> - how well does SQL-Server handles additions to VLTB
> (a real data warehouse is -never- static)
> - how well does it handle star schema implementations
> and the associated SQL queries
> - how well does it handle indexing on VLTB and what type
> of indexing techniques does it support (e.g. bitmap
> indexes)
> - how well does it handle full table scans
> - does it have parallel query support and how well is
> it implemented
>
> So once again, Microsoft is trying to blow a lot of smoke up our asses with
> their marketing hype. Well, I for one wet fart in their faces. Their
> TeraServer DOES NOT prove anything.
>
> Anyone wants to argue with me, please bring your lead pipe to
> comp.databases.oracle.server - that's where the Real Databasers (tm) hang
> out.
>
> As always, IMHO.
>
> regards,
> Billy
> PS. My opinion in flame bait format - so don't get on your high horses and
> please flame away. I would really like to know if SQL-Server will be able to
> handle a -real- terabyte database.
Received on Wed Jul 08 1998 - 00:00:00 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US