Enabling new, data-driven 5G services through a digital network twin

A TM Forum project is using a digital twin network (DTN) to demonstrate how telcos can build network twins to help manage the greater automation required by B2B2x and industrial 5G services. The proof of concept Catalyst project uses real-time data and knowledge from network operations, then creates an intelligent network to facilities network planning, construction, optimization and the autonomous operation of events.

The scale, complexity and dynamism of networks are increasing dramatically, as businesses strive to innovate with on-demand network slicing, low latency and higher capacity connectivity. Yet current telecom networks are not sufficiently digitized and intelligent to deliver next-generation network service performance and network operators are under pressure to cut operational costs.

More specifically:

• The radio access, transport and core networks are managed separately – these domains need to be coordinated for end-to-end assurance in real time.
• Network configuration is mostly done offline and requires manual interventions, making it slow and inflexible.
• There is no discrete management through closed-loop decision making, but low trial and error tolerance which is too slow and unwieldy.
• There is no service level agreement (SLA) assurance mechanism, which requires simulation, analysis, predictive and verification capabilities.

DTNs leverage AI in a virtual-real interactive closed-loop system centered on data and models. They help operators to realize the full lifecycle management of network planning, construction, operation, maintenance, and optimization, and facilitate digital transformation of the network. The communications service provider (CSP) is supported by participants AsiaInfo Technologies, Huawei Technologies, and SI-TECH Information Technology. China Mobile is looking to leverage the advances from this Catalyst in its commercial network, and set the project’s goals, and oversaw the development of technical specifications. The team drew on the work undertaken in TM Forum’s Autonomous Network Project, set up to help CSPs unlock over $700 billion of new revenues from industrial networks and B2B2x opportunities that require automation to manage the billions of devices connected to the internet as well as deliver on the ultra-low latency and high reliability many 5G use cases require. Virtual and physical network interactions The team chose the scenario of broadcasting major sports events to show end-to-end assurance of SLAs in a DTN to simulate the effect provisioning of an entire network, estimate the impact of the implementation, then tune the network configuration before deploying network services in the physical network. After the validation, the DTN will interact with the physical network to deliver the specific implementation plan for provisioning and modifying the network. AsiaInfo is responsible for developing the DTN Platform that will enable developers to create digital twins for end-to-end network elements. It is working with Huawei and SI-Tech to build digital twins for radio access, transport and core networks, as well as for the geographic information systems (GIS)and building information modelling (BIM) for the data center. The platform allows users to apply business rules, knowledge graphs and a machine learning inference model based on the particular use scenarios of network operation. It is built on top of a distributed data cloud, which provides the foundation of cloud native environment as well as aggregated data to digitalize the physical network entities. The developer uses an object-oriented approach to build a digital twin in the DTN Development Center, then establish business processes through interaction between the DTN and the Application Orchestration Center. Building a digital twin network Typically there are three steps to building DTN. The first is to model the physical network then digitalize it to create a visualization of physical equipment’s status and operation in real time.

The second step is to construct a knowledge graph of interactions between different network elements used to develop network simulation and orchestration. The third step is to train related AI models for specific DTNs based on different use scenarios and establish a DTN operation environment which covers different types of automated network operation. This shifts the network from being managed by human experts to AI-driven decision making. The team believes that network design based on DTNs will enable service providers to meet exactly the rapidly changing needs of businesses – and it will be much easier for network operation staff to simulate network performance based on historical data. The ultimate goal is to help operators expand the market for 5G deployments and applications in industry verticals while reducing network costs. Those using the services – industries and individuals – will benefit from a more stable network and more reliable, high quality services based on the comprehensive data, information, knowledge and intelligence provided by DTN. Watch this video to learn more about the Catalyst.