[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

900.0. "Benchmarks - defending" by DOD2::ROBERTS () Thu Mar 28 1991 20:55

    Does anyone have a good write-up on why TPC-B is a good benchmark? 
    Exactly what type of application work load does it test?  I currently
    have the results for Rdb/VMS and have recently read Oracle's results
    in the trade press.
    
    Why is TPC-B such a good benchmark.......
    
    lr

T.R Title User Personal
Name Date Lines

900.1 TPC-B is a standard BROKE::THOMAS Fri Mar 29 1991 20:51 13

T.R	Title	User	Personal Name	Date	Lines
900.1	TPC-B is a standard	BROKE::THOMAS		`Fri Mar 29 1991 20:51`	13
	TPC-B is an approved standardized benchmark, defined by the Transaction Processing Performance Council (TPC). As I recall (but I'm by no means certain) TPC-B is actually nothing more than a standardized TP1 benchmark. Therefore it is not a good benchmark. It does not test the database in a normal situation. It tests the database in a force-fed situation. How many transactions can you push through the database at one time. It does not consider any overhead involved with communicating with users or terminals, etc. The TCP-A benchmark is a much more reasonable benchmark since it tests the database in a more normal environment, including terminal communications, etc. In addition TPC is defining yet another benchmark, TPC-C, based on an order-entry application, which is designed to measure "business throughput". TPC-C has not yet been fully defined, but I expect that one day it will be the definitive benchmark to use.

    TPC-B is an approved standardized benchmark, defined by the Transaction
    Processing Performance Council (TPC).  As I recall (but I'm by no means 
    certain) TPC-B is actually nothing more than a standardized TP1 benchmark.  
    Therefore it is not a good benchmark.  It does not test the database in 
    a normal situation.  It tests the database in a force-fed situation.  How 
    many transactions can you push through the database at one time.  It does 
    not consider any overhead involved with communicating with users or 
    terminals, etc.  The TCP-A benchmark is a much more reasonable benchmark 
    since it tests the database in a more normal environment, including 
    terminal communications, etc.  In addition TPC is defining yet another 
    benchmark, TPC-C, based on an order-entry application, which is designed 
    to measure "business throughput".  TPC-C has not yet been fully defined, 
    but I expect that one day it will be the definitive benchmark to use.