> On 13 Jul 2021, at 10:52, Christer Holmberg <[email protected]> 
> wrote:
> 
> Hi,
> 
>>> When you say “does not support Outbound”, are you referring to the whole 
>>> mechanism, or to the specific 
>>> reuse-of-registration-connection-for-incoming-calls part?
>> 
>> We have had a lot of discussions about this in the past, where I focused on 
>> the latter. In the current set of standards a server is not allowed to reuse 
>> the incoming TLS connection for outbound requests. Only SIP outbound allows 
>> this. 
> 
> Correct.
> 
> However, many still do it, because in addition to the TLS issues it is also 
> needed for NAT traversal.
Exactly. Kamailio does that, because it’s the only way even if we break the 
RFCs.
> 
>> That’s why I started a discussion about a “half-outbound” - much like the 
>> use of Outbound in SIP over WebSockets today.
> 
> That specific part could be defined in a separate specification, similar to 
> what we did for SIP keep-alive: people wanted to have a mechanism for 
> negotiating keep-alives, without having to implement everything else in 
> Outbound.

Yes, that was my idea. But at the time it seemed like people either did not 
understand or did not see the need.
Maybe we have a different situation now. It is needed, as we need RFC-compliant 
security in SIP.

/O
> 
> Regards,
> 
> Christer
> 
> 
> 
> ________________________________________
> From: stir <mailto:[email protected]> on behalf of Olle E. Johansson 
> <mailto:[email protected]>
> Sent: Tuesday, July 13, 2021 9:32:56 AM
> To: Roman Shpount <mailto:[email protected]>
> Cc: mailto:[email protected] <mailto:[email protected]>; Mary Barnes 
> <mailto:[email protected]>; Salz, Rich 
> <mailto:[email protected]>; mailto:[email protected] 
> <mailto:[email protected]>
> Subject: Re: [stir] [Acme] http://NYTimes.com: How Do You Stop Robocalls? 
>  
> 
> 
> 
> 13 juli 2021 kl. 06:58 skrev Roman Shpount <mailto:[email protected]>:
> 
>  At the same time, SIP over TLS has many performance and reliability issues 
> that would need to be addressed before it is ready for industry-wide 
> deployment.
> 
> There’s also a lack of applicable standards for TLS usage, as I’ve pointed 
> out a few times, but the working group seems to have no energy to fix. SIP 
> over TLS from the SIP phone side requires implementation of SIP outbound, 
> which we never successfully tested at any SIPit. I know of a few 
> implementations now, but haven’t tested them together.
> 
> Made this presentation five years ago
> https://www.slideshare.net/oej/sip-tls-security-in-a-peer-to-peer-world
> 
> /O
> _______________________________________________
> stir mailing list
> mailto:[email protected]
> https://www.ietf.org/mailman/listinfo/stir
> 

_______________________________________________
Acme mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/acme

Reply via email to