Open collaboration platforms are key to autonomous networks
In advance of her participation in the Digital Transformation World Series virtual event, China Mobile's Lingli Deng explains how the service provider envisions partnering to enable autonomous driving networks.
07 Oct 2020
Open collaboration platforms are key to autonomous networks
China Mobile’s Lingli Deng will speak during the Digital Transformation World Series virtual event on Thursday, October 15, about why China Mobile is committed to promoting open source and standards collaboration around 5G and AI innovations. In advance, she explains in this blog how the company envisions partnering to enable autonomous driving networks.
During the last one to two years, the industry has been studying the phenomenon of autonomous driving networks. A series of white papers and work items have been published or are under development to specify what they are, their use cases and how to classify their levels. TM Forum’s first white paper on Autonomous Networks, published in May 2019 is a good example. (Editor’s note: TM Forum has just published a second Autonomous Networks white paper, which you can find here.)
From the highest level, the white paper highlights four parts. In addition to the definition of a series of autonomy levels and a collection of use cases, it provides a hierarchical architecture, in which it is recommended to evolve resource management into an autonomous domain. This is the basis for network and service automation on top and a methodology to prioritize the use cases for deployment by combining considerations from business drivers, technology maturity and coverage scale.
My Digital Transformation World Series presentation will explain three open issues for network operators to enable autonomous driving networks, including:
Transitioning from case-specific solutions in lower levels to cover all scenarios in higher levels
Lineardependence on technology and trust in the proposed hierarchical framework, which raises complexity and uncertainty
Use-case driven methodology because it is difficult to achieve convergence and arrive at a common strategy among network operators and service providers or to enable new cases in a sustainable and future-proof manner.
To address the remaining open issues for network autonomy, a three-stage path for network operators is needed to achieve the ultimate goal of full network autonomy:
Do the functional enhancement for specific high-priority cases at the beginning.
Stitch together end-to-end cases later.
Finally arrive at industrial convergence for the best outcome.
The key differentiators to ensure success in the last stages are to build and verify a common architecture across levels and standardize cross-layer interfaces, while in the earlier stages it is important to drive alignment among standards-development organizations. In particular, an existing open industrial platform as shown below provides an optimal way to explore, converge and verify cross-layer interfaces.
Last but not the least, it is critical to drive metric and benchmarking standards development for known use cases, and to build general DevOps tooling and management process for unknown use cases to come in the future.
Motivated by the vision for an open collaboration platform, our team at China Mobile has been working on relevant aspects in both open source and standards organizations. In particular, we are working in the following three sub-groups in Linux Foundation Networking (LFN):
The LFN End User Advisory Group (LFN-EUAG), which is an exclusive group of network operator representatives focusing on requirement mining for strategy convergence
Technical projects, including the Open Network Automation Platform (ONAP) to build a Reference Implementation for Common Functional Architecture, and
Open Verification Platform (OVP), the LFN’s automated testing and certification program, to provide DevOps tooling, benchmarking and certification procedures for network autonomy solutions.
In conclusion, the goal for network autonomy is defined in two dimensions: industrial convergence and building an open collaboration platform. Industrial convergence is the key for reducing the cost to both functional development and case coverage for any single vendor or single network operator. Building an open collaboration platform for cohesively developing both a reference implementation for the case-agnostic general functional architecture and standardized data/control or intelligence assistance exchanges via its external or internal interfaces would be the easy way for us to kick off and stay in the converged direction towards network autonomy.