Service Management and Orchestration (SMO) for OAM

This project focus on a docker-compose deployment solution for SMO/OAM Components.

Introduction

With respect to Operation and Maintenance (OAM) the SMO implements the O1-interface and OpenFronthaul Management-Plane consumers. According to the O-RAN OAM Architecture and the O-RAN OAM Interface Specification, the SMO implements a NETCONF Client for configuration and a HTTP/REST/VES server for receiving all kind of events in VES format.

The O-RAN-SC OAM deployment contains an OpenDaylight based NETCONF client and an ONAP VES Collector. Kafka is used as massage router for communication between the components. The Keycloak implementation offers and Identity service, while traefik acts as reverse proxy to terminate all incoming https traffic. For storing data in a persistence way the implementation of the mariaDB project is used.

SMO OAM Components

This docker-compose file starts a pre-configured, self-contained SDN-R solution with the following components.

• Identity ... representing an KeyCloak based identity service for centralized user management. Please note that the implementation does not support IPv6. Therefore, its own network is required called 'DMZ'.

• Controller single node instance

  ... representing the NETCONF consumer on the Service Management and Orchestration framework (SMO) for O-RAN O1 interface and/or O-RAN OpenFronthaul Management Plane and/or other NETCONF/YANG schemas implemented by the OpenDaylight project.

• VES collector

  ... representing the VES (REST) provider at SMO for all kind of events.

• Messages ... representing SMO MessageRouter component, includes message-router

• Gateway ... representing a revers proxy terminating TLS traffic (https, NETCONF)

Prerequisites

Resources

The solution was tested on a VM with

• 4x Core
• 16 GBit RAM
• 50 Gbit Storage

Operating (HOST) System

$ cat /etc/os-release | grep PRETTY_NAME
PRETTY_NAME="Ubuntu 24.04.1 LTS"

Docker

$ docker --version
Docker version 27.5.1, build 9f9e405

Please follow the required docker daemon configuration as documented in the following README.md:

• ./smo/common/docker/README.md

Docker Compose

$ docker compose version
Docker Compose version v2.32.4

GIT

$ git --version
git version 2.43.0

Python

$ python3 --version
Python 3.12.3

A python parser package is required.

python3 -m venv .oam
source .oam/bin/activate
pip3 install requirements.txt

ETC Host (DNS function)

Your local IP and your used interface is required. Please use the following script to modify all .env and other configuration files accordingly. The script will find automatically the interface and its ip-address to the internet. You can check its usage with option "-h":

python3 ./adopt_to_environment.py -h

Please run the script with your preferred fully qualified domain name, you like to use in your browser address bar.

python3 ./adopt_to_environment.py -d <your-smo-fqdn>

you can revert the settings in the modified .env and configuration files using the option "-r".

python3 ./adopt_to_environment.py -d <your-smo-fqdn> -r

It is beneficial (but not mandatory) adding the following line add the end of your ~/.bashrc file. I will suppress warnings when python script do not verify self signed certificates for HTTPS communication.

export PYTHONWARNINGS="ignore:Unverified HTTPS request"

Please modify the /etc/hosts of your system or modify the DNS of your environment.

• <your-system>: is the hostname of the system, where the browser is started

• <deployment-system-ipv4>: is the IP address of the system where the solution will be deployed

$ cat /etc/hosts
127.0.0.1	               localhost
127.0.1.1	               <your-system>

# SMO OAM development system
<deployment-system-ipv4>                   smo.o-ran-sc.org
<deployment-system-ipv4>           gateway.smo.o-ran-sc.org
<deployment-system-ipv4>          identity.smo.o-ran-sc.org
<deployment-system-ipv4>          messages.smo.o-ran-sc.org
<deployment-system-ipv4>      kafka-bridge.smo.o-ran-sc.org
<deployment-system-ipv4>          kafka-ui.smo.o-ran-sc.org
<deployment-system-ipv4>         odlux.oam.smo.o-ran-sc.org
<deployment-system-ipv4>         flows.oam.smo.o-ran-sc.org
<deployment-system-ipv4>         tests.oam.smo.o-ran-sc.org
<deployment-system-ipv4>    controller.dcn.smo.o-ran-sc.org
<deployment-system-ipv4> ves-collector.dcn.smo.o-ran-sc.org

Usage

Bring Up Solution

Short story

The following commands should be invoked. More detailed can be found in the next chapters.

source .oam/bin/activate
./setup.sh

Simulated network

Before starting the simulated network, you need to locally build the docker images. This is because of copyright issues with the 3GPP YANG models.

The build should be pretty straightforward. The repository containing the PyNTS code needs to be cloned and then a command needs to be ran for building the images. Run this from another terminal, in another folder, not in this repo.

git clone "https://gerrit.o-ran-sc.org/r/sim/o1-ofhmp-interfaces"
cd o1-ofhmp-interfaces
make build-all

After everything is built successfully, you can return to your solution folder here and start the network.

docker compose -f network/docker-compose.yaml up -d
docker compose -f network/docker-compose.yaml restart pynts-o-du-o1

Check (adjust if required) environment variables

nano smo/common/.env
nano smo/oam/.env
nano network/.env

Startup solution

Please note that it is necessary to configure first the identity service, before starting further docker images.

The several docker-compose yaml files must be started in the right order as listed below:

docker compose -f infra/docker-compose.yaml up -d
docker compose -f smo/common/docker-compose.yaml up -d
python smo/common/identity/config.py

The python script configure the users within the identity service (keycloak). A system user (%USER) is also created with administration rights.

docker compose -f smo/oam/docker-compose.yaml up -d

Looking into the ONAP SDN-R logs will give you the startup procedure.

docker logs -f controller

If you see the login page (https://odlux.oam.smo.o-ran-sc.org) you are good to go and can start the (simulated) network.

docker compose -f network/docker-compose.yaml up -d

Usually the first ves:event gets lost. Please restart the O-DU docker container(s) to send a second ves:pnfRegistration.

docker compose -f network/docker-compose.yaml restart pynts-o-du-o1

The simulated O-DU and O-RUs are pre-configured according to O-RAN hybrid architecture.

O-RU - NETCONF Call HOME and NETCONF notifications O-DU - ves:pnfRegistration and ves:fault, ves:heartbeat

SDN-R reads the fault events from DMaaP and processes them. Finally the fault events are visible in ODLUX.

Log files and karaf console

ODL karaf.logs

docker exec -it controller tail -f /opt/opendaylight/data/log/karaf.log

ves-collector logs

docker logs -f ves-collector

Customizing Solution

'.env' file contains customizing parameters

Verification Solution

Access to SDN-R ODLUX

https://odlux.oam.smo.o-ran-sc.org

User: admin 

Password: // see .env file

In case of trouble, please update the commands with your customized '.env' file.

Access to Node Red Flows

https://flows.oam.smo.o-ran-sc.org

User: admin 

Password: // see .env file

In case of trouble, please update the commands with your customized '.env' file.

Terminate solution

To stop all container please respect the following order

docker compose -f network/docker-compose.yaml down
docker compose -f smo/apps/docker-compose.yaml down
docker compose -f smo/oam/docker-compose.yaml down
docker compose -f smo/common/docker-compose.yaml down
docker compose -f infra/docker-compose.yaml down

alternative:

./teardown.sh

Cleanup

!!! be careful if other stopped containers are on the same system

docker system prune -a -f

Troubleshooting

In most cases the .env setting do not fit to the environment and need to be adjusted.

Please make sure that the network settings to not overlap with other networks.

The commands ...

docker ps -a
docker compose ps
docker rm -f $(docker ps -aq)