Re: [stir] I-D Action: draft-kaplan-stir-ikes-out-00.txt

[Torrey Searle <[e-mail]>]

Hello,

Regarding the SIP <->XMPP inter-working of SUBSCRIBE.  Subscriptions in SIP
are time limited, and last forever in XMPP, as a result, sip<->xmpp gateway
may need to generate several SIP subscribes in response to a single XMPP
subscribe.  Perhaps it's worthwhile to  note that only the Dialog Creating
and Dialog Destroying subscriptions will have a signature, and the rest of
the in-dialog requests should be considered trusted as well?


Regards,
Torrey


On Tue, Jul 16, 2013 at 5:19 AM, Hadriel Kaplan
<[e-mail]>wrote:

>
> On Jul 15, 2013, at 4:05 PM, Torrey Searle <[e-mail]> wrote:
>
> > I really like your draft, especially the fact that inter networks with
> ss7.  Just have a initial comment that in the case of the UUI header, the
> spec should probably specify that  the Protocol Discriminator for the UUI
> header should be set to 00 - User Specific Coding.  Though it might me an
> interesting question if it is possible to use a new value for the protocol
> discriminator to easily identify that the value in the UUI header is a
> signature.
>
> Crap, I forgot about the protocol discriminator.  I don't mean I forgot to
> mention it, I mean I forgot about the byte it takes, not to mention the
> type and length bytes.  That means there're only 128 bytes available, which
> for a 1024-bit private key means all of those 128 bytes will be the
> signature.  So I'll have to move the key index and timestamp fields into
> the Call-Reference param instead.  Ugh.
>
> But anyway, yeah good catch, the discriminator should probably be 0x00.
>
>
> > Also how about the case where [e-mail] gets aliased to an e164
> when reaching the pstn gateway?  I assume the pstn gateway would "own" the
> e164 and can re-sign the call before forwarding, but would it be
> interesting to mention this case in the spec?
>
> Does that ever happen in the PSTN gateways?  I know it happens in some
> service providers (like skype for example), but I thought it happened on
> some SIP or H.323 system just before it reached the PSTN GW.  Regardless,
> yes I should mention that in the draft too.
>
> Thanks for the feedback!
>
> -hadriel
>
>