Re: [dispatch] draft-jesske-sipping-etsi-ngn-reason-04

[<[e-mail]>]

Adam wrote:
> My point is that increasing the amount of ISUP information that we 
> include in SIP message headers is going down a slippery slope, the 
> ultimate conclusion of which is including SIP header fields for every 
> ISUP parameter.
This is a valid point, but the proposed alternative has severe problems.
If the reason header is the only piece of ISUP information we need,
carrying the whole ISUP message is not really elegant. SIP application
servers (AS) would have to implement ISUP parsing just to check a simple
value. The content of a Reason header could be checked and used by any
plain AS. Do we really want to perpetuate ISUP code in SIP networks?

> We have a means for sending ISUP fields transparently through a SIP 
> network (RFC 3398). I think it's folly to revisit that set of
decisions 
> simply because someone wants a less-general point solution.
We were talking about scenarios where the PSTN gateway does not know
whether a call will be terminated by a different PSTN gateway, or by a
SIP user agent. So the originating gateway would have to encapsulate
ISUP in SIP, and SIP user agents would receive it. This would violate
regulatory requirements in various countries. As S/MIME is not feasible,
we would have to introduce yet another B2BUA just to filter ISUP from
SIP messages for end user devices.


Wolfgang