The role of intent APIs in cost-effective autonomous networks

Telecommunications networks are crucial to digital transformation of enterprise, yet CSPs face many barriers to meeting the diverse needs of different industries. Complex network operations and maintenance (O&M) and high OPEX requirements – ranging from initial telecom cloud infrastructure to full-service-enabled network services – are just some of the challenges to contend with. CSP involvement in enterprise transformation is therefore characterized at times – at least, currently – as siloed, inflexible, expensive to maintain, and with limited vertical integration.

Reducing the resources needed for digital transformation is key to mass deployment. As a result, the industry has been working on fundamentally improving O&M capabilities and implementing a highly stable and self-healing autonomous network architecture. The Intent-driven autonomous networks - Phase II Catalyst, which focused on intent-based APIs, is crucial to realizing this goal.

Intent APIs: an industry-wide project

Intent APIs are based around adaptive requirements-based communication, suitable for achieving separation of the concerns that enable AI-native automation. This means telling the network which goals we want to achieve, but relying on its ability to independently determine how they are delivered and maintained. In this Catalyst, the project team used intent APIs and intent-driven operations to balance the conflicting needs of customers with service provider business goals. The Catalyst demonstrates that intent-driven, multi-layer, autonomous operations can be achieved with TM Forum Open APIs which underpin lifecycle management, domain-specific requirement expression and intent management function.

Early applications and scenarios of intent APIs

The project team explored how these APIs could be applied to business scenarios, and by designing intent interfaces of the business layer, service layer and resource layer, they developed a number of innovative use cases. One notable example was dynamic pricing of connectivity services. In this scenario, CSPs use idle bandwidth to provide competitive high-quality connection services for price-sensitive or QoS-sensitive users or third-party partners. This application integrates two scenarios: intent-driven service ordering and service assurance.

Based on the autonomous network reference architecture defined in TMF IG1251, this application provides self-ordering, self-configuration, and self-service assurance capabilities through hierarchical autonomy – enabling zero-wait, zero-touch, and zero-fault experience for businesses and services. Customers’ service ordering intention is translated layer by layer into configuration instructions that can be executed by network elements. After each network element correctly executes the configuration instructions, services are provisioned. Customers’ QoS intentions are translated into service and resource assurance policies, and SLAs can be flexibly guaranteed and optimized through real-time SLA awareness.

In demonstrating how intent APIs can be practised in a manner that is suitable for numerous industries, CSPs have taken a step closer to realizing autonomous networks that serve the needs of a broad range of vertical industries. They have also laid the path for next steps in defining new interfaces to provide advanced automation and intelligence capabilities – both of which will be critical in implementing more efficient collaboration and interaction with users and third parties, and enabling agile service innovation and differentiated user experience.

Project leaders explain the value of intent APIs

Kevin McDonnell (Catalyst Lead & Senior Director, Huawei) says intent-driven autonomy goes well beyond simple automation: “Manual operations are ‘do this’, automated operations are ‘if this, then do this’ - but intent-driven operations are ‘I want this!’. This is a much more fundamental shift in the way requirements are expressed, and it’s central to allowing AI-native systems to find more optimal solutions.” Massimo Banzi (Standardization Manager, TIM) says that “the key benefit of the intent-driven approach is the decoupling of Intent (‘what’) from the implementation, the closed-loops, as this will increase interoperability and vendor independence across autonomous systems”.

Lester Thomas (Chief Systems Architect, Vodafone) examines the benefits of intent-driven using a historical context: “As our systems become more complex, it will become increasingly difficult for humans to create optimized policies – this is where the intent-driven approach comes in, helping optimize both end-user experience and operator revenues. It will shift 5G from one-size-fits-all to a network that can be tailored to meet diverse user needs.” Intent enables that potential to account both for heterogeneous needs and rapidly increasingly network complexity – making it a powerful approach that with a far broader impact than any singular technology of the moment could achieve.