| 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 |
Hello world -
we need some advice concerning INGRES performance vs. Rdb.
A partner (PSI, a CSO in Berlin) started developing an application using
INGRES and INGRES-Tools (4gl, Forms) on a VAX 6000-310 (VMS).
They are experiencing severe performance problems.
For comparison reasons, they did a little benchmark (two simple insert
statements) using Rdb vs Ingres (no 4gl, just plain SQL). In this simple
comparison Rdb performed 5 times faster than INGRES.
As they were forced by theit customer to use INGRES in this project,
they now need advice if Rdb can be considered faster on VMS than INGRES.
If so, they would try to convince their customer to switch over to Rdb.
Well, I know a comparison depends on the DB-Design, the kind of
application, configuration etc. What I would like yo know: are there
some simple, rule of thump figures comparing INGRES vs. Rdb performance?
Also, it would be interesting to here about conversions of a INGRES-Tools
application from INGRES-DB to Rdb.
If Rdb is faster than INGRES in generall, does this hold true in a
INGRES-Tools environment?
OK - any hints, informations, experience on this matter would be highly
appreciated!
Andreas
| T.R | Title | User | Personal Name | Date | Lines |
|---|---|---|---|---|---|
| 960.1 | LOOK IN INGRES_TOOLS_FOR_RDB | TRCOA::MCMULLEN | Ken McMullen | Mon Jul 22 1991 21:33 | 1 |
Also look in NOVA::INGRES_TOOLS_FOR_RDB | |||||
| 960.2 | already done... | BER::MOENCH | Andreas Moench @BEO, Berlin | Tue Jul 23 1991 09:42 | 8 |
re.-1 >> Also look in NOVA::INGRES_TOOLS_FOR_RDB Sorry, I forgott to mention: this topic was crossposted there... Thanks for the hint! Andreas | |||||
| 960.3 | INGRES vs Rdb Performance | COOKIE::RDOWNS | Tue Jul 30 1991 00:56 | 16 | |
I have done a very thorough comparison of INGRES vs ORACLE vs Rdb on
the 6310 for the "Order-Entry Workload", and the results show INGRES to
be very inferior to ORACLE on the OLTP portion of the workload while
ORACLE is itself inferior to Rdb. For the End-User computing (EUC)
portion of the workload, INGRES is very inferior to both Rdb and
ORACLE, which are about equal. The numbers and complete results are in
a report titled "INGRES, ORACLE and Rdb, ORDER ENTRY WORKLOAD TEST
RESULTS" by Rudolph Downs and Francois Raab. This report is now in
distribution. For a copy , please send mail to COOKIE::MCMINN.
Thanks,
Rudy Downs
COOKIE::RDOWNS
CXN1/6 Colorado Springs, USA.
| |||||
| 960.4 | WHO301::BOWERS | Dave Bowers @WHO | Fri Nov 01 1991 22:15 | 11 | |
My personal experience has been that INGRES is slower than Rdb, and especially for queries where the effort of parsing the SQL (which they do at runtime) becomes significant vs. the ttime to retrieve the data. INGRES tools on Rdb take a serious performance hit because of the need to replicate in large part the INGRES metadata. This causes metadata references to be a lot more complex than the Rdb equivalent, since they need to go through both their stuff and ours to get to the data. I've also seen problems caused by their inability to deal with simple deadlock and lock conflict errors. -dave | |||||