John Merrells <[EMAIL PROTECTED]> writes:

> On 11-Feb-06, at 3:17 PM, Eric Rescorla wrote:
>
>> Method of ticket validation
>> This draft validates the ticket by having the Membersite send a digest
>> to the Homesite and get an ACK. It's not clear why this is desirable.
>> Wouldn't it be simpler to have the Homesite digitally sign the ticket
>> (the key could be delivered in the initial capabilities discovery
>> phase) and then let the Membersite do the verification directly?
>> I appreciate that there's a freshness concern, but this can
>> be alleviated using the usual nonce-based anti-replay techniques.
>
> The motivation wasn't freshness. The dix:/message-id parameter
> is a nonce that takes care of this.
>
> The motivation was to get all the binary crypto code out of the MS to
> ease adoption. We learnt from our prior experience with the SXIP
> protocol that this was a barrier to adoption. Writing good DSIG code
> for all platforms/stacks/languages is tedious and expensive and worse
> increases the number of lines of code that a MS developer has to
> write to enable a site. [SXIP 1.0 worked this way.]

What do you mean "binary crypto code"? You've got a hash algorithm,
no? At worst, you could share a pairwise secret between the MS and
the HS during the initial discovery phase and use that to key
a MAC. (This is of course only safe if you're doing that exchange
over SSL/TLS, but that's true of your scheme too...)

Anyway, I don't really find this that convincing. Java certainly
comes with built-in public key functionality and there are modules
for Python, Perl and PHP (it's actually a compilation flag for PHP).
Yes, it's not zero effort, but it's not exactly prohibitive either.

-Ekr



_______________________________________________
dix mailing list
[email protected]
https://www1.ietf.org/mailman/listinfo/dix

Reply via email to