[Dnssec-deployment] "NSEC3 Hash Performance" by Yuri Schaeffer

[Stephane Bortzmeyer]

On Wed, Jul 21, 2010 at 03:15:41PM +0000,
 Paul Vixie <vixie at isc.org> wrote 
 a message of 11 lines which said:

> seen on the iepg list today (thank you jaap!)
> 
> > http://www.nlnetlabs.nl/downloads/publications/nsec3_hash_performance.pdf

The opinion I sent to the author:

But I am disappointed by the choice of tested values. You test
(section 2.1) at regular intervals while I would be more interested by
values around 1. Noone will sign with 100 or 1000 iterations (1 for
.ORG, 10 for .GOV or .COM.BR, etc).

I would prefer to see systematic testing from 1 to 10 iterations and
then, if you want (although I don't think there is a practical
interest), samples from 10 to high values.

In the same way, I would prefer a log horizontal scale for figures 1
and 2. Your paper does not tell me what it costs to choose 1, 3 or 5
iterations.