[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference ulysse::rdb_vms_competition

Title:	DEC Rdb against the World

Moderator:	HERON::GODFRIND

Created:	Fri Jun 12 1987
Last Modified:	Thu Feb 23 1995
Last Successful Update:	Fri Jun 06 1997
Number of topics:	1348
Total number of notes:	5438

753.0. "Sizing INGRES 6.2, 6.3" by TRCA01::LEW () Fri Sep 28 1990 17:07

I have a client who is running INGRES 6.2 on VMS 4.7 and an 8530.
They are experiencing severe performance problems and
would like to upgrade their hardware.  The maximum user load
is 65 but this will be increasing in the near future.

Their application is written using INGRES's 4gl ABF.

I scanned the Notes file and was able to come up with the
information for older versions of INGRES (notes 79 and 257). 

What I need is some up to date information on the current products.

Any help would be much appreciated.


1) Can anyone recommend any "rules of thumb" with respect to :

	- number of users per VUP.

	- memory requirements per INGRES user

2) Does anyone have any horror stories or warnings that I should
   be aware of specifically in the areas of

	- INGRES and Clusters, SMP

	- INGRES and large user loads


Thanks........... Ivan

T.R	Title	User	Personal Name	Date	Lines
753.1	Here's one story	KIER::KIER	My grandson is the NRA!	`Fri Sep 28 1990 18:53`	28
	Your milage may vary... I have a customer for whom I did a performance analysis on Ingres V5 which they have used as a benchmark against their V6.2 conversion. Remember this is their particular application and may not apply to your circumstance: Users: 118 average - peaks of 140 Three-node cluster: 8550, 8600, 6320 - ~17 VUPs Application written in Fortran with Embedded QUEL All tables are on three-spindle shadow-set volumes - spread across four shadow sets - RA81s and RA82s Ingres V5 - .052 VUP/user 1.3 MB/user Ingres V6.2 - .062 VUP/user 2.5 MB/user (yeah, the new client/server version requires MORE memory, at least for this application) They do not use the fast commit stuff, using a full cluster implementation. They were in severe pain until they were shipped an emergency order of memory and now performance is back within 10-20% of where it was with V5. Mike
753.2		TROA02::NAISH	SQL4ME Paul Naish DTN 631-7280	`Mon Oct 01 1990 04:58`	22
	Ivan, your best source of help at TRC is Dave Vanduyvenvoorde DTN 637-3475. He has been working with INGRES and ULTRX/SQL. A couple of other things. You didn't indicate how the need for a hardware upgrade was determined. It could be they need a good DB tuning. We also sell capacity planning services which could be used to effectively predict H/W requirements based on current loading. Note, that this assume well behaved application, DB and system software. If they do need new hardware look to ULTRIX. ULTRIX/SQL is now bundled at no charge (the engine, not tools) and is based on INGRES V6.2 . See conference SMURF::ULTRIX_SQL for more info. It is also important that Digital not provide sizing information on its own for 3rd party products. Why? If the configuration does not perform as promised, only Digital is on the hook for making it work. Have the sales rep involve the INGRES for help. Get what ever they recommend in writing. This is somewhat different in the scope of a capacity plan. There, you need to work the issue of scalability with the vendor. One other minor issue, if they do upgrade their system, they will have to upgrade to VMS V5.4, a non-small step!