Autonomous, intent-driven networks cannot be achieved without open network data

FREE PERSPECTIVE

The ability to rapidly change and improve processes based on business needs and use cases will be crucial to operating networks in the 5G era. Traditional network data collection systems predominantly use proprietary, non-standard data formats and models to capture and process data. This restricts the agility and operations of communications service providers (CSPs) because they have to pre- and post-process network data and data scientists have to produce customised models for different use cases to derive insights and support operational processes. Openness must be embraced in the network to empower higher-level applications and enable automation.

Standardised access to network data enables CSPs to perform in-depth analytics and to generate insights in order to create efficiencies and accelerate their journeys towards autonomous networks. To achieve this, CSPs must support the following three distinct capabilities.

• Network data openness enables CSPs to democratise FCAPS data that comes from the network using standard content and encoding formats, so that it can be turned into actionable insights.

• Use case-based APIs enable CSPs to use the curated data for higher-layer network automation, including the automation of network planning, engineering and roll-out processes, thereby significantly reducing network opex and time to market.

• Intent-driven applications powered by the underlying network intelligence enable industry digital transformation by exposing the network services to enterprises via open APIs.

Standard northbound open APIs are the key to fully exploiting open network data. The industry is coalescing around the open APIs NETCONF and RESTCONF, as well as the YANG models for network automation use cases. The open APIs can abstract the complexity of the underlying network from the management functions, thereby resulting in lean operations using a standardised, intent-driven approach, thereby enabling autonomous networks.