Abstract

The OpenDaylight project is an open source platform for Software Defined Networking (SDN) that uses open protocols to provide centralized, programmatic control and network device monitoring. It allows automating networks of any size and scale. PANTHEON.tech, a TOP open-source contributor, actively participates in the LINUX FOUNDATION NETWORKING environment. Our contributions play a crucial role in shaping the OpenDaylight project. In this article, we delve into the insights of our developers, exploring their perspectives on the OpenDaylight environment and its practical applications.

How important is OpenDaylight for your clients?

Our customers operate modern complex and multi-layered network environments. They face a huge and demanding challenge to integrate many different systems, frameworks, engines, tools, technologies, etc. The OpenDaylight environment is a key enabler to help them automate this integration. We use a variety of OpenDaylight APIs to do this.

Which API do you use the most?

The RESTCONF REST APIs provide a convenient way for the customer’s external systems, applications and scripts to access the data in the OpenDaylight datastore or within the mounted devices. REST is also used to allow external applications to trigger various processes/functions within OpenDaylight using RPC calls. OpenAPI has also been introduced to facilitate this type of REST interaction. JMX APIs are another way for external applications to interact with OpenDaylight, although this is mostly used for monitoring, statistics gathering and health checks. Some customers integrate JMX APIs with their own messaging systems to publish alarm notifications to various external nodes.

How do you deal with backups and disaster recovery?

OpenDaylight is primarily used as a network device connector and configurator. Thanks to its internal data store (Model-driven Service Abstraction Layer MD-SAL), it’s also used as a disaster recovery and backup tool – so in cases where some of the devices die and need to be rebooted cleanly, we use OpenDaylight to restore them to their last known state.

What protocols do you use?

The main protocol used to connect and communicate with devices is NETCONF. Other protocols are also available, but with varying levels of production quality. Connecting and communicating with devices is usually done in 2 ways:

Directly – by calling the NETCONF APIs to create Netty-based NETCONF sessions with the devices. These sessions are then used by custom applications.
Via mount points – accessing the internal state of the device.

How do you secure communication?

To secure NETCONF communication, OpenDaylight comes with an AAA (Authentication, Authorisation, Accounting) plugin. However, some clients have their own security policies and procedures. For this purpose, the original AAA can be replaced by custom plugins to better meet the client’s requirements.

What are OpenDaylight’s clustering capabilities good for?

OpenDaylight is often used for high availability, fault tolerance and geo-redundancy. This is achieved using either OpenDaylight’s clustering capabilities, Daexim utilities or other hybrid solutions.

Customers can customize their OpenDaylight deployments – the number of OpenDaylight instances (members) that form the cluster, their voting rights and the distribution of data between them. Customers often split data into different shards and then decide which shards to persist and which members to host them on – adding high availability.

Can you cluster multiple OpenDaylight instances?

Some customers prefer to run multiple standalone OpenDaylight instances and connect them into an HA cluster via a higher level orchestrator. This can be particularly useful in cases where load balancing is an important part of their deployment and needs to be carefully managed.

img pahEPhg0xUWn3pW7STuMS