[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

18.0. "DECUS_France benchmark" by VIRGO::RINKUS () Mon Jul 27 1987 21:45

    Can anyone help me obtain a copy of the DECUS-France benchmark 
    between RDB, Oracle and Ingres. I had a copy but it was so good
    
    someone decided to keep it for thenselves; noW I can't recall
    where I got it in the first place!!!
    
    			
    Thanks for your help.
    
    				-paul r.
T.RTitleUserPersonal
Name		DateLines
18.1I Have the AnswerCEDSWS::BOOTHA career of MISunderstandingMon Jul 27 1987 23:384
      Look in the NOVA::RDB notes file. That will tell you the location
    of the OLD_RDB notes file. I think the benchmarke results were
    somewhere around note 1116.
    --- Michael Booth
18.2Here you go !!OTOO01::WARKTue Jul 28 1987 22:043
    Check your mail, Paul.
    
    	regards, Bill from Ottawa
18.3I made it !PRSNRD::GROSGURINWed Aug 05 1987 18:4054
Hello,

My name is Jean-Michel GROSGURIN, and I recently heard about this
notesfile. 

I work in DIGITAL France at Puteaux as SWAS-spec since October, 1986. 

The DECUS-France benchmark was conducted in MARCH, 1987 and I can
give any information about it as long as I performed it myself...

Moreover, note that :

1. Ingres did not attended the benchmark,

2. Oracle did, but did not officially published the results.
   
   These results (we had them...) have actually been put in the 
   BANZAI::RDB_OLD previous notesfile in a summarized form.

The versions used were Rdb V2.2 and Oracle 5.17.

The same benchmark has been conducted last year (ie. MARCH, 1986),  
and at that time Rdb V2.0 had been used.

It's interesting to note that in some tests based on the elapsed
time, in conflicting access situations, the results with Rdb 2.2 
were 3 times better than with Rdb 2.0, mainly because of the
I/O mechanism improvements.

I have a rather complete description of all the tests run for 
this benchmark, and also of other tests run afterwards with Rdb 
2.2 only. In 2 sentences, the results are the following [] :

[  On a VAX 8700, with a 1Gb database spread on 3 RA81,
   we reached 4 TPS with an average elapsed time of 2.5 seconds,
   an I/O throuput of 70 I/O per second, 20 database users and
   80% of the CPU used.

   NB: the "T" of TPS means "average order processing transaction", 
       relating only to the database request, with no TDMS or FMS 
       load at all.
    
       In the DECUS-France benchmark, the transactions were sort of
       "atomic database transactions", therefore they were far less 
       costly in terms of CPU and I/O.                                ]


I wrote the whole report in French; if anyone is interested I can
either forward it, or try to extract some valuable information 
with some comments in English.


Regards,				
			Jean-Michel