[WG-IMS] Scope of IMS in OpenSIPS - RFC

[Bogdan-Andrei Iancu]

Hi Giovanni,

Thanks for this valuable summary on the http2/diameter in 5g/4g - things 
are much, much clear now :)

Best regards,

Bogdan-Andrei Iancu

OpenSIPS Founder and Developer
   https://www.opensips-solutions.com
   https://www.siphub.com

On 06.12.2023 15:19, Giovanni Maruzzelli wrote:
> On Tue, Dec 5, 2023 at 11:23 AM Bogdan-Andrei Iancu 
> <bogdan at opensips.org> wrote:
>
>     I did some reading about this NSA vs SA and as I understand, this
>     will not impact the actual functionalities we need, but more the
>     logic and interfacing with other components (HSS or PC(R)F).
>
>
> Maybe I gave the impression it is a problem of formal requirements 
> (and that's not), actually in my view it is a problem of (open source) 
> implementations.
>
> One of the problems is that in 5G (let's call it NR, NewRadio), 
> Open5GS implemented a different interface in PCF vs the one 
> implemented in PCRF.
>
> So, while in LTE (4G) Open5GS uses Diameter for interfacing, in NR 
> (5G) it has only implemented HTTP2.
> Then, we need HTTP2 and NR specific logic for implementing VoNR (Voice 
> over New Radio) using Open5GS as NR core.
>
>     More or less, from OpenSIPS perspective, we need the same set of
>     features, but the scripting will differ (as logic and DIAMETER
>     messages).
>
>
> Yep, definitely: if all specs would have been implemented, VoNR would 
> have *both* diameter and http2, so, it would be a superset of VoLTE 
> with added http2 and some more things, and you can decide to interface 
> it from IMS (that's us) via diameter (like in VoLTE) or via HTTP2 
> (VoNR specific). But Open5GS does not implement diameter in NR.
>
> Yes, as Carlsten said, we may implement a network function of a 
> translator between diameter and http2, no idea if that is easier than 
> implementing http2 into (x)CSCF.
> Maybe (just maybe) with kubernetes and all these deployment 
> environments mainly tailored to web, they will heavily push with http2 
> for additional services.
> Or maybe in near future Open5GS will implement diameter in NR.
>
> Without Open5GS interfacing us (IMS)  we'll not be able to actually 
> test real implementation with actual 5G phones (and devil is always in 
> details, simulations are nice, but...)
>
> -giovanni
> --
> Sincerely,
>
> Giovanni Maruzzelli
> OpenTelecom.IT
> cell: +39 347 266 56 18
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opensips.org/pipermail/wg-ims/attachments/20231206/b4387d68/attachment.html>