Future Communication Testbed

Motivation

The Future Communication Testbed (FCT) at DAI Labor serves as a platform for the validation and demonstration of its projects on current research topics where cutting-edge communication network infrastructure plays a key role.

Objective

The FCT aims at providing a playground for latest network research activities and concepts that require realistic environments, as well as providing a technological foundation for a wide range of application domains. The application domains currently handled by the FCT include Autonomous Driving, eHealth, Smart City, and Smart Grid.

Technology

The FCT consists of several hierarchical building blocks that highly utilizes virtualization technologies and can be flexibly configured for the application scenario at hand, allowing for the synergies of different research projects to materialize on the joint setup. The FCT components currently in use and under continuous development are:

Mobile Edge Computing Infrastructure – Distributed compute infrastructure deployed on a road stretch of 3.5km in downtown Berlin, in close vicinity to sensors and devices, to support the real-time data analysis of mobile and stationary sensors close to the network edge. Each MEC unit features an industrial-grade GPU computing platform.

Flexible Transport Network – SDN-based transport network that can be dynamically adapted and configured to meet the requirements of the device layer and application domain. A microwave backhaul supports the high-rate interconnection of the MEC infrastructure with the DAI-Labor data center, in addition to the use of C-V2X and DSRC links for vehicular connections.

Lightweight 5G Core Network – Implements a lightweight set of 5G Core Network functionality to manage devices and the virtualized network infrastructure.

AI-driven Network Service Orchestrator – Based on Machine Learning approaches, the FCT Network Service Orchestrator extends the management and orchestration of the lightweight 5G Core Network with intelligent and adaptive behavior.


The testbed is currently being used and extended in the following projects: