Unlocking Autonomous Networks with Intent-Based Automation
By Yuval Stein, AVP Technologies, TEOCO
The logic which underpins ‘intent’ in autonomous networks is highly compelling. By using artificial intelligence (AI), network orchestration and machine learning (ML) algorithms to automate routine functions, the network then becomes more intelligent, more autonomous, and more adept in terms of adapting to changing workloads. This new networking paradigm provides the ability to design a more flexible and scalable network, thereby delivering a superlative performance.
The motivation behind the development of this concept lies in the fact that business, service, and network operations are becoming increasingly complex as services become more dynamic, fragmented and distributed. At the same time, while the advancement of 5G brings with it the promise of more lucrative revenues streams for network operators as well as differentiated services that better meet customer’s exacting needs, it too will add to the overall complexity. However, this growing network intricacy is outpacing operators’ abilities to oversee, control and assure customer experience, service operations, cost and network performance via the use of traditional methods and tools.
Looking ahead, manual and static programming and rules-based automation needs to evolve more towards model and knowledge-based approaches formed on the basis of the intent behind business, service and resource requirements and constraints. By switching to this type of approach, services can be transformed and developed more autonomously as business strategies, goals and customer requirements change.
The future sustainability of network operations
Autonomous Networks (AN) provide operators with the opportunity to actualize meaningful service quality improvements and also unlock signficnant revenue growth. While the latest technological advancements in 5G, AI, IoT, cloud and edge computing have been adopted globally (and are notably being leveraged by key enterprise verticals in support of their specialized business needs), the digital transformation of networks has been necessary to remain in lockstep with this growing network complexity. As operators continue the roll-out of their 5G networks the manual and labour-intensive methods traditionally employed to manage the network and services are no longer sustainable. Given the extraordinary pressures that network operators are under, such forces will make it harder for them to maintain their competitive advantage and continue to deliver high quality services to their customers.
The implementation of full automation is the only viable option to overcome the challenges associated with the dynamicity of modern networks. The arrival of closed-loop network automation that is self-healing and self-scaling will result in: greater network stability and performance; the near complete elimination of human intervention; a simplified service design; improved service parameters; the ability to immediately react to network events as they occur to avoid outages.
One of the promising new trends in relation to AN today is the concept of intent-based management. An intent-based model promotes greater separation of tasks and clarifies roles and responsibilities between the “owner” of the intent and the “handler” of that same intent. In this scenario, intent is the expression of the requirements that an AN party needs to meet, making it a key concept in the establishment of an advanced, highly capable, intelligent network. Intent will typically be passed between different management domains enabling each one of them to work autonomously.
The importance of Autonomous Networks to operations
ANs are critical to operating complex and dynamic cloud-native networks efficiently, ensuring that operators remain aligned with their customers’ requirements and continue to meet their evolving expectations. ANs will enable operators to handle the complexities of orchestrating network resources to enable rapid service innovation. They will also be able to harmonize the assurance and orchestration resources so that the whole network is unified and synchronized in its autonomy through closed-loop automation.
ML and AI are also key enablers of the closed-loop automation concept that is intrinsic to autonomous networks. These intelligent algorithms can handle the complex decision making required to configure and optimize network resources and to orchestrate services so that changes to the network are implemented in ways that are not disruptive to other areas of the network. Looking ahead, the goal of ML and AI will also be to enable predictive operations: algorithms will learn to spot patterns emerging from network service degradations and employ automated remediation routines to adjust the network parameters with the overall aim of reducing down time and increasing the quality of service.
Intent-based management compliments ML and AI technologies within the context of autonomous networks. It helps to separate the ownership of management goals from implementation details, so the actors that need to understand the “what”, do not need to be concerned with or understand the “how”. By creating a language that is straightforward for both owners and handlers to understand, it creates the space for new standards to emerge i.e., the ability to express intent, thereby creating an open ecosystem that reduces the dependency on specific vendors, opens up the market for new, disruptive players, fosters innovation and reduces costs.
Aspirational autonomous networks
Today, the technologies needed to create highly capable AN are just within reach. The efficient management of the extensive and increasingly complex networks that operators must grapple with – made more intricate and convoluted by the arrival of 5G and its associated technology ecosystems – will require the right capabilities that are designed to adapt to new situations and changing needs without human intervention. This is where AN with intent-based management can make a profound contribution, helping network operators to optimize their networks with efficiency and minimalism—and in the end, achieve more with less.