Mark Andrews wrote:
> In message <[email protected]>, P Vixie writes:
>> > Tsig won't scale for something like this. Please consider sig0.
>
> I've got no objection to sig(0) but why won't it scale?  There is
> a existing relationship so public key cyptography isn't needed.

sneaker net for key management, including revocation, emergency key
swaps, for delegation-mostly domains that might have tens of millions of
different subdomain operators, is a recipe for disaster.

> Sig(0) would require the KEY record to be in the parent zone or to
> be held by the registrar in a seperate database.  In the later case
> you either need a database of KEY records or a database of TSIG
> keys.  As far as I can tell there is no difference in the scaling
> requirements.

the KEY RR would also be SIG(0) updateable.

>
> Sig(0) might be marginally more secure as only one side holds
> material than needs to be kept private.

take that marginal difference and multiply it by six billion to get the
internet wide impact over time.

-- 
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