> Fred Morris <mailto:[email protected]>
> Friday, November 28, 2014 3:07 PM
>
> ... is not mathematically necessary. As a simple counterexample, XOR is
> commutative and associative: it doesn't matter the order you XOR multiple
> blocks in. Not saying XOR is the One True Way, just that implementation
> details like that are probably a distraction at this point.

any zone-level signature has to be crypto-authentic. XOR is too easy to
"fix up", as in, add or delete your desired changes, compare the new
checksum to the old one, then add a TXT RR that causes the new checksum
to match the old one.

so, i'm not in favour of zone-level signatures per se, but if they're
coming, then marka@isc's characterization of them as "sorting and
hashing" is mathematically nec'y.

-- 
Paul Vixie
_______________________________________________
dns-operations mailing list
[email protected]
https://lists.dns-oarc.net/mailman/listinfo/dns-operations
dns-jobs mailing list
https://lists.dns-oarc.net/mailman/listinfo/dns-jobs

Reply via email to