Re: [Sidrops] WG Adoption call for draft-harrison-sidrops-manifest-numbers-01 - ENDS 04/15/2024 (April 15 2024)

[Rob Austein <[e-mail]>]

I support adoption because it's worth discussing.  I'm not yet
convinced that there's a real problem that needs solving here.

On Tue, 02 Apr 2024 12:35:36 -0400, Job Snijders wrote:
...
> 1) The algorithm presented by RFC 1982 ("Serial Number Arithmetic") has
>    a significant shortcoming: there are sequence numbers for which
>    comparison is undefined.

For any given serial number there's exactly one value for which
comparison is undefined.  Given the size of the numbers here, you're
(much) more likely to be struck by lightening before reading the end
of this sentence than you are to hit that case by accident, so if you
ever do hit it you can assume it's an attack. :)

>    I consider arithmetic operations on a number space this large an
>    unwelcome complication to be avoided if possible.

Implementations I've seen use bignums to hold these numbers, and the
computational expense of one additional bignum subtraction would be
lost in the noise compared to the signature check.