[29339] in Kerberos
Re: kerberos capacity planning questions
daemon@ATHENA.MIT.EDU (John Hascall)
Wed Feb 20 17:24:28 2008
To: kerberos@mit.edu
In-reply-to: Your message of Wed, 20 Feb 2008 12:51:10 -0800.
<87ablve4ht.fsf@windlord.stanford.edu>
Date: Wed, 20 Feb 2008 16:23:11 CST
Message-ID: <5876.1203546191@malison.ait.iastate.edu>
From: John Hascall <john@iastate.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
Errors-To: kerberos-bounces@mit.edu
Russ Allbery <rra@stanford.edu> writes:
> Tim Metz <tpmetz@ucdavis.edu> writes:
> > Greetings,
> > As part of a campus working group currently evaluating Kerberos product
> > options, I have been tasked with researching the following two
> > requirements for MIT Kerberos:
> > (1) Must house more than 200,000 accounts.
> > (2) Must issue more than 3 tickets per second.
> MIT Kerberos can trivially satisfy both of those with any reasonable
> server hardware. We were doing more authentications per second than that
> on a now-obsolete Dell 1750. We're now running the primary Kerberos
> server, which handles almost all the authentications on a Dell 1950.
Even on non-reasonable hardware. In 1990 I tested our
pathetic-by-current-standards DECstation 2100-based KDC
at around 40 tickets/sec (single DES then, of course).
Our peak second so far today (also on a Dell 1950)
was 58 tickets issued (we'll do around a million
over the course of the whole day).
(FWIW: peak minute was 1248, peak hour 66274)
John
________________________________________________
Kerberos mailing list Kerberos@mit.edu
https://mailman.mit.edu/mailman/listinfo/kerberos
