Telstra revolutionizes network service automation with ‘Knowledge Plane’ innovation

Who: Telstra

What: Developed a groundbreaking approach to autonomous networks using an in-house knowledge graph and “Knowledge Plane” to encode network knowledge and enable more intelligent, dynamic network operations

How: Building on its ODA-based reference architecture including TM Forum’s best practices for deploying NaaS and digital twins, Telstra invented a new way to model relationships between network components and encode knowledge, allowing the company to move beyond scripted network service automation to an intent-driven approach.

Results:

• Capability to process a higher volume of orders
• Processing time reduced by 30%
• Service creation and onboarding time reduced from 12-18 months to 3 months
• Increased Net Promoter Score (NPS)

Australian operator Telstra has invented a revolutionary approach to network service automation that the company intends to share with the telecoms industry through TM Forum. At the core of the innovation is the concept of a “Knowledge Plane” that fundamentally changes how engineers’ knowledge of network domains (such as, IP and optical transport networks, the radio access network or the 5G core network) is encoded and utilized.

The best way to explain the innovation is to consider the “before” and “after” of network service automation, says Telstra Chief Architect Mark Sanders.

“Before, in order to automate a part of the network, people would write automation scripts,” he explains. “You had to think of every use case that could occur.” So, achieving outcomes like managing capacity or energy usage or avoiding failure scenarios all had to be coded into the script.

But such an approach doesn’t scale, especially when Telstra and other network operators are aiming for autonomous networks. Based on TM Forum’s definition, highly autonomous networks – deemed Level 4 – mark the transition between traditional automation of human-defined process behavior and autonomous behavior, where systems make decisions independent of humans.

“What we’ve done is fundamentally different,” Sanders says. “The knowledge around what the network domain is and how the resources function is encoded in knowledge graphs right up front. That’s the big shift.”

Focusing on intent

By providing this new way of encoding engineers’ knowledge of the network rather than just point-in-time data, the Knowledge Plane enables operations teams to think in terms of outcomes, or intents, instead of step-by-step processes. Importantly, the Knowledge Plane spans multiple domains – or “entities” as Telstra calls them – allowing for the exchange of knowledge between them and more efficient, optimized operations end to end across the network.

To build the knowledge graphs and develop the Knowledge Plane, Telstra had to create a telco-specific ontology, meaning a bespoke way of encoding network knowledge. Since none existed, the company created an ontology using the TM Forum Information Framework. Telstra also created a Domain Specific Language (DSL) for interacting with the knowledge graphs and a graph-aware application that empowers a streamlined orchestrator.

Another key component of the solution is a network differencing engine, which compares current service states with desired outcomes. This allows for precise, intent-based actions that optimize network performance and align with business objectives, making the network effectively self-healing.

Matt Beanland, Principal Engineer at Telstra, who invented the Network Differencing Engine concept and led the implementation, uses a rally car analogy: “The model-driven differencing engine is running as the navigator and providing turn-by-turn instructions based on your goals, policy and experience. The simplified orchestrator is the driver strangely enough – not the navigator – executing the instructions and reporting their effects. Together they are a high-performance team.”

The idea is to be able “tune and refactor the way that services are provided – ‘the how’ – dynamically, without doing any kind of big project across the telco or without any analysis of downstream dependencies,” Beanland adds. “That’s really where we’re headed.”

Part of broader transformation

Telstra’s innovative approach to network service automation is the result of a much broader transformation effort. The company achieved running on ODA status a year ago and is continuing to leverage TM Forum’s Open Digital Architecture (ODA), including the Open APIs, to develop its Telstra Reference Architecture Model (TRAM).

“One of our great learnings out of this is that if we’re going to fundamentally change how we work, it goes deeper than just thinking about APIs and the ODA,” says Sanders. “It’s about how we change our skills and change our approach.”

He adds: “We had to unlearn behaviors and approach designing networks differently. We had to approach building products differently. We had to think of how we engage our customers differently.”

Indeed, it took a while to get the Knowledge Plane concept off the ground. Telstra developed it about a year and a half ago and began using it in the fixed-access network. Then, it was extended to the company’s Enterprise Connectivity Services, which demand complex orchestration across multiple domains.

“We had been doing orchestration and automation in a certain way for a long period of time,” Sanders says. “We had to change our thinking. And we’re continuing to do that.”

To build the Knowledge Plane, Telstra drew on TM Forum’s NaaS Transformation Guide and used the NaaS API Component Suite. The resulting solution integrates Telstra’s NaaS gateway, incorporating multiple Open APIs.

The company also drew inspiration from a document called Guiding Principles for Building and Measuring Autonomous Network Solutions, which was influential in shaping Telstra’s approach to the principles around observability and responsibility. These are important for delineating machine versus human responsibilities and ensuring interoperability.

Finally, TM Forum's assets focusing on digital twins were also helpful. The Knowledge Plane essentially mirrors the live network service, acting as a virtual representation of network services, encapsulating various configurations, roles, relationships and dependencies.

Business impacts

The business impact of the new approach has been significant. Service creation and onboarding time has dropped from 12-18 months to 3 months. And with the capability to process a higher volume of orders per month, Telstra has also been able to reduce order processing time by 30%.

For customers, this backend innovation has translated into a better experience, evident in an increase in Telstra’s Net Promoter Score (NPS).

Now, the company is looking to launch a new capability that will allow customers to design their own network services through an intuitive interface – a level of flexibility and control that was unthinkable with legacy systems.

Industry-wide innovation

But Sanders and his team are perhaps most excited about the impact the Knowledge Plane concept could have on the wider telecoms industry.

“We have the idea that the industry should progress this concept more deeply – and the ontology that might go with it,” Sanders says. “I think the interest will be there… In some sense I think we’re a little bit ahead of our time, meaning that we have to articulate why this is a good idea.”

He adds: “I think it's important to introduce it. Our goal is not just to make this for Telstra, but to cast the net wider for the industry. That’s absolutely our goal.”