5G-MEDIA MAPE service

Introduction

The 5G-MEDIA MAPE service, as part of the 5G-MEDIA Service Virtualization Platform, is responsible to:

• collect the original monitoring data from NFVIs, VNFs/NSs (applications), QoEs etc
• correlate the original monitoring data with the running VNFs and NS and adapt them in a uniform structure
• store the adapted monitoring data after the adaptation process
• visualize the monitoring data in a dashboard
• analyze the performance of the running NSs/VNFs and suggest optimization actions that improve the QoΕ in the end-user
• apply the optimization actions interacting with th appropriate service in the 5G-MEDIA Service Virtualization Platform

The 5G-MEDIA MAPE service consists of a set of components:

• Translator
• Cache (Redis has been used)
• Data-importer
• Database (InfluxDB has been used)
• Dashboard (Grafana has been used)
• Data lifecycle manager
• Accounting agent
• Optimization components
  • O-CNO-predictive-optimizer
  • O-CNO-arbitrator
  • SS-CNO for Tele-Immesrive games based on reinforcement learning (UC1)
  • SS-CNO for Mobile Contribution (UC2)
  • SS-CNO for Remote and Smart Production based on reinforcement learning (UC2)
  • SS-CNO for anomaly detection over vCDN services based on deeplearning (UC3)
  • SS-CNO for anomaly detection over vCDN services based on supervised learning (UC3)
• Executor
• Recommender
• Logging server (Grayload has been used)
• Issue tracking server (Redmine has been used)

The communication and the messages exchange among the above components is achieved through the 5G-MEDIA publish/subscribe broker. The Apache Kafka has been selected as publish/subscribe broker. Follow the instructions of the repository to build the docker image including the kafka broker / zookeeper and deploy it through the docker-compose. Moreover, Kafka manager is used to visualize statistics for the kafka broker.

The NFVI monitoring data are sent in the publish/subscribe broker through the below three publishers and then are processed by the Translator:

• OpenStack-kafka-publisher
• Kubernetes-kafka-publisher
• OpenNebula-kafka-publisher

The deployment of the MAPE is done as a project using the docker-compose. A docker image is built for every component (unless it is available in a docker registry).

The structure of this repository is:

• The deployment folder includes the deployment and configuration files.
• The scripts folder includes various bash script files.
• The figures folder includes the figures of the repository.

Deployment

Deploy the services:

$ sudo docker-compose -f docker-compose-mape.yml up -d

Deploy the services with multiple translator instances. Indicative example:

$ sudo docker-compose -f docker-compose-mape-ncsrd.yml up -d --scale translator=3

List of the docker container including the healthcheck label:

$ sudo docker ps -a

Stop and remove the services:

$ sudo docker-compose -f docker-compose-mape.yml down

Stop and remove the services, remove the unused volumes:

$ sudo docker-compose -f docker-compose-mape.yml down --volumes

Check the volumes list:

$ sudo docker volume ls

Inspect a specific volume (e.g. the grafana volume):

$ sudo docker volume inspect <volume_name>

Check the existing networks:

$ sudo docker network ls

Check the dedicated network for the MAPE services:

$ sudo docker network inspect <network_name>

Get the logs of the involved services:

 $ sudo docker-compose -f docker-compose-mape.yml logs

Get the logs of the involved services in real time:

 $ sudo docker-compose -f docker-compose-mape.yml logs -f

Get the logs of a specific container (e.g. grafana):

$ sudo docker logs -f mape-grafana

Network settings

The traffic in a set of ports must be allowed from the firewalls:

• 80/tcp (grafana)
• 8080/tcp (kafka manager)
• range 8880-8888/tcp (services under supervisor)
• 6379/tcp (redis)
• 8086/tcp (influxDB)

In case that you include the ceilometer-kafka-publisher, you must allow traffic in: 10000/udp.

Test

Ensure that docker engine has been installed and is running:

$ sudo service docker status

Check the docker engine version:

$ sudo docker --version

Docker version 18.03.0-ce, build 0520e24

Authors

• Singular Logic pathanasoulis@ep.singularlogic.eu
• UCL

Contributors

• Contact with Authors

Acknowledgements

This project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 761699. The dissemination of results herein reflects only the author’s view and the European Commission is not responsible for any use that may be made of the information it contains.

License

Apache 2.0