dbench scalability testing

[David Collier-Brown]

Anton Blanchard wrote:
Real netbench runs will be useful now because tridge tells me
smbtorture
has not been verified to approximate netbench at the higher end of the
scale.

	I had a chance to do one set of runs on a large
	system used for NetBench testing with 48 clients.
	The smbtorture curve had a different maximum
	but a very similar shape, offset by a constant
	term.  This implies that smbtorture is 
	weakly predictive in the mathematical sense,
	and a valid simulation of the NetBench load
	in the practical sense. 

	I suspect we can approximate NetBench with
		N = (a * S) + c
	for some value of a and c 

	a and c will vary somewhat with site/configuration

--dave
-- 
David Collier-Brown,           | Always do right. This will gratify 
Performance & Engineering Team | some people and astonish the rest.
Americas Customer Engineering  |                      -- Mark Twain
(905) 415-2849                 | davecb at canada.sun.com