Closed gruyaume closed 1 month ago
Short answer is yes
Long answer: This is something that we had discussed in the last 2-3 TST meetings and it is part of the roadmap. However, for the implementation, we are going to start with the SBI for the 3GPP release upgrade and after that we will move to work on the N4 interface (PFCP). Please take a look at the "Agenda & Meeting Notes" for the link to the roadmap. Is this an effort you guys can help with?
I would have to align with the team, but yes, this should be an effort we can help with.
We are already using this in UPF and we wish to move SMF also to this go-pfcp module. ITs on roadmap and we are happy if you could contribute to this. Thanks
Description
I'm opening this issue to entertain the possibility of deprecating this repository. There is already a well maintained GO PFCP library out there, wmnsk/go-pfcp, and the SMF could handle PFCP communication using it instead of
omec-project/pfcp
. Here I want to discuss whether or not this is a good idea.Rationale
Ease of life for developers
We are currently using two different projects for PFCP communication between the SMF and the UPF making it harder to understand and contribute changes involving PFCP communication. The UPF uses
wmnsk/go-pfcp
and the SMF usesomec-project/pfcp
, both with the same purpose. Why not depend on the 1 project for PFCP communication instead of 2.Security and reliability
go-pfcp
is well maintained, tracks changes to 3GPP releases, has 53 dependents (making it more likely to receive bug fixes and security patches), and has only 1 dependency (go-cmp).Maintainability
Archiving this repository would reduce by 1 the total number of repositories maintained for Aether, making the project easier to support with a limited amount of developers.
Effort
This activity would need to be done in three phases:
wmnsk/go-pfcp
in the SMFomec-project/pfcp
to maintenance mode where only security patches would be done if requiredomec-project/pfcp
after X amount of time