Re: [dispatch] New draft: draft-holmberg-dispatch-received-realm-00

[Christer Holmberg <[e-mail]>]

Hi Peter,

Thanks for your comments! See inline.

> I agree that it is useful for a (transit) network to know which network a request arrived from, as this
> draft argues. But isn't this information to be in the Via header field already in the entry before the
> transit network? The draft mentions that inter-operator identifier cannot be used but I think the
>draft should expand on why a new header field parameter is needed.

Via headers may contain IP addresses, and they cannot always be used to identify the network. Also, the consumer of the information would have to scan through the Via headers, and figure out which belongs to the consumer network and which belongs to the adjacent network.

The purpose of the identifier is to have an explicit indicator, which prevents such scanning etc.

I try to explain that in the draft, but maybe more text is needed :)

Regards,

Christer


From: dispatch [mailto:[e-mail]] On Behalf Of Christer Holmberg
Sent: 08 June 2015 20:15
To: [e-mail]<mailto:[e-mail]>
Subject: [dispatch] New draft: draft-holmberg-dispatch-received-realm-00

Hi,

I've submitted a new draft to dispatch: draft-holmberg-dispatch-received-realm-00. The draft defines a new Via header field parameter, "received-realm", which network entry entities can insert in order to inform other entities in the network from which network the request was received.

Previous versions of the draft have been submitted to SIPCORE, but I was advised by the SIPCORE chairs to submit it to DISPATCH.

Compared to previous versions, there is a technical change: the entity which inserts a parameter also calculates a HMAC value, which is inserted together with the network value.

Regards,

Christer