On Thu, Jul 17, 2008 at 04:02:34AM -0400, Rob Austein wrote:
> > To put it in a slightly different way, what I wondering was is there a
> > means of specifying this "minimal expression" you refer to?
> 
> Aye, that is the question.  I've tried to specify an algorithm for
> this twice already, and failed both times.

Unless I have missed something about the subtleties of an ROA, the
prefixes contained therein are of the form:
<prefix> [maxlen <n>] | n >= prefix-length.

Given any two prefixes, A and B of the above form where A has the
shorter prefix length:
If prefix A covers prefix B and
maxlength A >= maxlength B
Then A completely covers B and B is redundant.

If prefix A covers prefix B and
maxlength B > maxlength A
Then B has a portion of it's address space that is not covered by A.

Given the above, you can pair-wise compare each prefix and eliminate
redundant prefixes.  (Insert common optimizations of pre-sorting by
prefix length, etc.)

I'd honestly just expect implementors to toss the prefixes into a prefix
trie and not worry about the redundant nodes.

> As Jeff and the other Robs have been discussing, there's some question
> of whether one actually needs an equivalence operation; more
> precisely, whether one needs an efficient equivalence operation.

The main concern would be someone's ROA having so much noise in it that
it's consider a resource exhaustion attack.  Given that people would be
using these things to say "hey, I want traffic to reach my network!" it
seems less than likely that someone would intentionally try to do that,
especially since the response is to blacklist the ROA.

-- Jeff
_______________________________________________
Sidr mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/sidr

Reply via email to