Hello Francesco,

I agree with Brett, be wary of fashion fades.

Voip is """" just """" routing around packets, by inspecting and modifying them

This needs to be completely aware of all states of all things (transactions, dialogs, application, user, billing, cdr, etc) at all moments for all calls (and if you need to land specific calls on specific media boxes - eg freeswitches - then you need even to know the relashionships between an incoming call and all the legs already in the whole platform)

On top of that, there are all the particularity of using real time streaming in udp, instead of a transactional query/answer model like db or http (and smtp, etc)

So, all in all, for the core part of voip, you are well served by kamailio+freeswitch

You may want to optimize it, but that is the way it has to be

Obviously, the applications on the back of voip can be refactored (or completely rewritten) to be microservices (the GUIs, the CRMs, etc) if there's gain in doing that

Have a great day!!

answered from mobile, please pardon terseness and typos,
-giovanni

On Tue, Mar 25, 2025, 17:44 Francesco via sr-users <sr-users@lists.kamailio.org> wrote:

Dear all,

I hope this email finds you well. I am reaching out to gather insights and suggestions regarding a potential redesign of our VoIP system to improve scalability, ease of management, and upgradeability.

Current Setup:

  • Our system is not multi-tenant and is based on Kamailio and FreeSWITCH in a distributed architecture (as example of distributed architecture, see: https://ufdcimages.uflib.ufl.edu/UF/E0/04/92/89/00001/THOMPSON_C.pdf).
  • Everything runs inside LXC containers across multiple nodes.
  • One of the challenges we face is OS upgrades, which are cumbersome and require significant manual intervention.
  • At the moment the code for voip is done with ansible

Given these challenges, I am considering whether a microservices-based architecture could be a viable option. However, I have some concerns regarding:

  • SIP and RTP handling with NATed containers (e.g., Kubernetes, Docker Swarm) and potential issues with media traversal.
  • Performance impact when shifting real-time traffic handling to containerized environments.
  • Upgrade and deployment strategies—how to achieve smoother, rolling upgrades without service interruptions.
  • Alternative approaches that might be more effective, such as a hybrid model combining LXC with containerized microservices for specific functions.

If anyone has experience transitioning a VoIP system to microservices or has insights on best practices for scalable, maintainable VoIP architectures, I would greatly appreciate your thoughts. What worked for you? What pitfalls should we be aware of?Do you have any documentation that I can check?

Looking forward to learning from your experiences.

Thank you.

.: Francesco Colista

.: Francesco Colista
     


                                                        


__________________________________________________________
Kamailio - Users Mailing List - Non Commercial Discussions -- sr-users@lists.kamailio.org
To unsubscribe send an email to sr-users-leave@lists.kamailio.org
Important: keep the mailing list in the recipients, do not reply only to the sender!