< Previous by Date Date Index Next by Date >
< Previous in Thread Thread Index Next in Thread >

Re: [reSIProcate] RFC 5923 alias



On 11/04/12 14:55, Scott Godin wrote:
> Hi Daniel,
> 
> If you can use RFC5626 support, then RFC5923 is not required.  There is
> more functionality gained with RFC5626 use (ie. doesn't require TLS
> transport, provides a heartbeat mechanism, etc.).  I'm also not sure
> offhand if Polycom supports RFC5626 or not.  However if it doesn't then

I've checked: Polycom confirm that reg-id (RFC5626) is not supported in
the current firmware.

I'd be quite happy just to use RFC5923 with the restriction that it is
requires TLS, as I want TLS anyway.  I find that without TLS, there is a
risk of SIP-aware routers mangling the SIP packets, and that conflicts
with ICE.

I notice some of the Microsoft variations are supported (epid,
proxy=replace type stuff)

> you can try out the flow token mode of repro - this allows RFC5626 like
> behavior for endpoints that don't support it:

I tried that, I'm not 100% sure about the way these settings interact:
if I do EnableFlowTokens=true, then does the setting of DisableOutbound
make any difference?

How about the relationship between ClientNatDetectionMode and
EnableFlowTokens?  Does EnableFlowTokens supercede
ClientNatDetectionMode, or do they work together in some way?

> # Enable use of flow-tokens in non-outbound cases
> # WARNING: Before enabling this, ensure you have a RecordRouteUri setup,
> or are using
> # the alternate transport specification mechanism and defining a
> RecordRouteUri per
> # transport: TransportXRecordRouteUri
> EnableFlowTokens = false
> 
> It should be noted that some edge functionality of SIP will be broken
> with this mode.  Byron has documented this in InteropHelper.hxx:
>       // .bwc. If this is enabled, we will record-route with flow tokens 
>       // whenever possible. This will make things work with endpoints
> that don't
>       // use NAT traversal tricks. However, this will break several things:
>       // 1) Target-refreshes won't work.
>       // 2) Proxies that do not record-route may be implicitly included
> in the
>       //    route-set by this proxy, because a flow token may point to them.
>       // 3) Third-party registrations won't work.
> 
> Scott
> 
> On Tue, Apr 10, 2012 at 7:37 PM, Daniel Pocock <daniel@xxxxxxxxxxxxx
> <mailto:daniel@xxxxxxxxxxxxx>> wrote:
> 
> 
> 
> 
>     I'm just trying repro with a Polycom device that is behind NAT,
>     using TLS
> 
>     When the device tries to register, I notice that repro can't send the
>     407 challenge back to the device over the TLS connection.  The log just
>     says `discarding stray response: SipResp: 407 tid=8b5....'
> 
>     The Polycom device appears to support the `alias' attribute in the Via
>     header, for re-use of the TLS connection:
>     http://tools.ietf.org/html/rfc5923
> 
>     repro supports RFC 5626 (reg-id in the Contact header), but I'm not sure
>     if the Polycom handset supports that
> 
>     Can anyone comment on either of these solutions, i.e. is 5923
>     appropriate, has anyone got support for it in a branch somewhere
>     perhaps?
> 
>     _______________________________________________
>     resiprocate-devel mailing list
>     resiprocate-devel@xxxxxxxxxxxxxxx
>     <mailto:resiprocate-devel@xxxxxxxxxxxxxxx>
>     https://list.resiprocate.org/mailman/listinfo/resiprocate-devel
> 
>