5G EVE - MSO-LO interface

MSO-LO is a REST API application to provide an ETSI SOL 005 v2.6.1 compliant interface towards various NFV orchestrators.

OpenAPI specification is provided on Swagger Hub or in openapi directory.

The software is developed under the 5G EVE project.

List of authors/contributors:

• Francesco Lombardo, CNIT
• Matteo Pergolesi, CNIT
• Grzesik Michal, Orange
• Panek Grzegorz, Orange
• Chabiera Michal, Orange

Software is distributed under Apache License, Version 2.0

Install guide

Environment with iwf repository

If iwf-repository is available in your environment, edit docker-compose.yaml and change the relevant environment variables:

x-environment: &environment
  IWFREPO: 'true'
  IWFREPO_HTTPS: 'false'
  IWFREPO_HOST: '192.168.18.14'
  IWFREPO_PORT: '8087'
  IWFREPO_INTERVAL: '300'

Then, deploy with:

docker-compose up

Environment with local database

Edit docker-compose.yaml and disable iwf repository support.

x-environment: &environment
  IWFREPO: 'false'

Deploy with:

docker-compose up

Redis logical database mapping

In the following table we report all the Redis database indexes used in MSO-LO application and the relative purpose.

Index	Purpose
0	Celery Background Task
1	NS last operationState
2	OSM token cache

Simple test

You can test the app with:

curl --request GET --url http://127.0.0.1:80/nfvo

Uninstall

To remove the containers and volumes, use:

docker-compose down --remove-orphans --volumes

Developer guide

Branching model

There are two main branches:

• master: used for software releases, push not allowed
• development: used for daily work on code

To add a new feature, we follow this pattern:

1. Move to development branch: git checkout development
2. Create a new branch named after the feature you want to add. E.g.: git checkout -b onap_driver
3. Work freely on the branch
4. When done, merge your branch into development: git checkout development; git merge --no-ff onap_driver;

Note: The --no-ff option is useful to create a commit dedicated to the merge and record the existence of the feature branch. To add it to your git configuration so it is used by default (locally for this repo), use:

git config --add merge.ff false

Environment setup

For dependencies we use Pipenv.

Copy the mock files (needed for a correct build):

cd adaptation-layer/seed
cp nfvo_mock.json nfvo.json
cp nfvo_credentials_mock.json nfvo_credentials.json
cp rano_mock.json rano.json
cp rano_credentials_mock.json rano_credentials.json

To setup the environment use:

git checkout development
cd adaptation_layer
pipenv install --dev
# Create database with mock data
export FLASK_APP=.
export FLASK_ENV=development
pipenv run flask db upgrade
pipenv run flask seed
# Run the flask app
pipenv run flask run

Some features like notifications need celery and redis. Simply setup a docker container with redis and run a celery worker.

docker run -p 6379:6379 --name some-redis -d redis
export REDIS_HOST=localhost
celery -A tasks worker -B --loglevel=info

A docker-compose.dev.yml is also available. Remember to copy the mock files as said above for a correct build. Deploy with:

docker-compose -f docker-compose.dev.yml up --build

---

Please, always use pipenv to add/remove dependencies:

pipenv install <package-name>

pipenv uninstall <package-name>

If everything works with the new dependencies, run pipenv lock and commit both Pipfile and Pipfile.lock.

Add a new NFVO/RANO driver

To create a new NFVO/RANO driver, it is enough to create a new python module extending the Driver interface. For example, let's create adaptation_layer/driver/onap.py:

from .interface import Driver

class ONAP(Driver):
    pass

Now you can start implementing the methods contained in the Driver interface. Your IDE should suggest you to create stubs for methods to be overriden.

To enable a newly created driver, edit manager.py. The get_driver() method is simply a switch that returns an instance of the proper driver.

Add unit tests for a NFVO/RANO driver

In order to test our software against an NBI, we need to mock it. For this purpose, we use Prism. You can control the kind of HTTP response returned by Prism by modifying the request URL. Example: /nfvo/nfvo_osm1/ns?__code=200

To add unit tests for a driver, create a new python file in adaptation_layer/tests. Please refer to test_osm.py for examples.

Error handling

When throwing exceptions, please use the ones defined in error_handler.py. This allows the application to correctly create the corresponding response and status code.

Tests

Unit Tests

Unit tests can be executed by using Docker Compose files. The following unit tests are currently available:

• docker-compose.test-nfvo.yml Test NFVO information retrieve
• docker-compose.test-osm.yml Test interactions with a mocked OSM
• docker-compose.test-onap.yml Test interactions with a mocked ONAP
• docker-compose.test-ever.yml Test interactions with a mocked EVER

Example:

docker-compose --file docker-compose.test-osm.yml --project-name test-osm build
docker-compose --file docker-compose.test-osm.yml --project-name test-osm up --abort-on-container-exit --exit-code-from test-osm

Note: the --project-name parameter is necessary to distinguish test executions.

The file will run two containers:

1. A Prism server mocking the OSM NBI
2. A python container to execute unit tests

Unit tests execution for a new driver can be added by copying and modifying docker-compose.test-osm.yml.

Integration tests

Integration tests are run with Robot Framework. Please refer to the specific README.