Implementing Intent-Based Networking Without Lock-In, Part 2

The basics of intent-based networking and service automation—and why open source is the best way to implement them.

Part two of a two-part blog series.


In my last blog, Implementing Intent-Based Networking Without Lock-In, Part 1, I reviewed the basics of intent-based networking and service automation—what it is, why it matters, why open-source is the best way to implement it. Now, in the second part of this series, let’s take a closer look at Lumina’s open source solution to the problem. As you’ll see, it is entirely possible to move to intent-based service automation without locking yourself into a proprietary ecosystem. It’s also possible to start benefiting from intent-based automation right now—while building a foundation for even greater innovation in the future.


Implementing Open Source Intent-Based Networking

Lumina offers a lock-in-free approach to intent-based networking based on the OpenDaylight SDN Controller, the industry’s most popular open source controller. We couple this with our own innovations, which are upstreamed back to the OpenDaylight community.

Lumina’s open source Intent-Based Service Automation solution draws on:

  • Lumina SDN Controller (LSC): Lumina SDN Controller, based on the Linux Foundation-backed OpenDaylight project, is the industry’s leading and most widely deployed open source SDN controller. OpenDaylight has proven its mettle as a mature open source platform, backed by a broad, deep, and active community. The native controller platform offers a dynamic automation framework that operates based on data models expressed in YANG. The framework makes it possible to express intent using YANG models, and then translate that intent into specific network configurations in a multi-vendor environment.


  • Lumina Extension and Adaptation Platform (LEAP): LEAP takes intent-based networking to the next level by adding a framework for a rich set of scalable microservices around OpenDaylight. Using this extension framework, it becomes easy to integrate more sophisticated applications that can understand and process intent, such as a Natural Language Processing module. (While the industry is not yet deploying such radical innovations, the combination of LEAP and LSC makes such ideas possible.)

Together, this framework gives CSPs two key capabilities to translate intent into network action:

  • The ability to express intents in a vendor-neutral method that’s widely accepted and adopted by the industry
  • The ability to add different types of intent translations in a non-programmatic way, so that operations personnel can continually, dynamically add new intents as the need arises


Assuring Intent

Of course, CSP networks aren’t static—especially in the realm of 5G networks and services, which are highly dynamic. Once an abstraction framework expresses the high-level intent and configures it in the network, how do you make sure the network continues to deliver on that intent, even as conditions change? This is where service assurance systems come into play, and where you want policy-driven feedback loops that can take automated actions whenever the network deviates from the intent.

Lumina’s LEAP framework makes it possible to realize such closed-loop-feedback systems. The framework supports a rich set of southbound telemetry interfaces that can be used to monitor a variety of parameters, which can be summarized using analytics databases. CSPs can plug applications into this framework that can query these databases and invoke feedback loop calls back to the controller’s intent APIs.

In this way, CSPs gain a mechanism to assure automated actions in intent-driven networks. At the same time, they gain the freedom to continually update and change the underlying network infrastructure, without having to manually recode automation from scratch.


Making Open Source Intent-Based Networking a Reality

Like all Lumina software, our intent translation system is built as an in-house innovation but contributed back to the OpenDaylight community. You can find it under the component name ODL-Plastic. Plastic offers a rich framework to translate anything from northbound to southbound, be it a trivial data translation, or a more complex mapping that completely morphs one model into another. The framework provides a means of declaratively describing intent and how it should be translated. This enables field-deployable changes, making it a really powerful framework.

Bottom line: you can start implementing intent-based networking automation, and you don’t need to lock yourself into a third-party framework you don’t control to do it.

Additional Resources: