Programmable Infrastructures

A 6G network will overcome the static partitioning of functionality (in RAN/access/core etc.) and provide functions where they are actually needed. The available HW resources will be taken into account here. Application as well as network functions (e.g. game server vs. AAA) will be included and a continuous, agile adaptation to changing load/service requirements such as infrastructure availability will be considered. This results in an “organically growing” decentralized, robust, scalable – and thus evolutionary – 6G system. This vision of such an “organic infrastructure” requires, among other things, the ability to mediate between different variants of a functionality to combine them into beneficial overall services, while supporting the requirements of the data path. Functions can range from very lean and stateless to complex-state, heterogeneously connected into services, with different management requirements (e.g., launching new instances vs. state migration).

Decisions can also be decentralized. Deeper integration of (classical) radio, core, and user functions also opens rich data sources for AI/ML procedures involved in data path processes (“linerate inference”) as well as management processes (e.g., reconfiguration of services when changing handover strategy). For the development and operation of such a system, which is subject to continuous evolution, the ideas of DevOps and web-scale software engineering must be developed further into a “network scale DevOps” (e.g. parallel operation of different versions, “continuous fault injection” for robustness).

Project goals

  • Development of an organic core network that is able to:
    • Dynamically reassemble itself according to service requirements and available infrastructure.
    • Dynamically migrate to the available network infrastructure as needed
    • Provide customized connectivity services (mobility and QoS support, authentication and authorization, etc.)
    • Support automated network management – ultra-flexible scaling and “seamless” software upgrades
  • The solution is based on:
    • comprehensive consideration of the benefits of software engineering technologies for connectivity services
    • Adoption of software-based architecture (SBA) through to user equipment (UE)
    • Dynamic integration of heterogeneous backhauls
    • Deployment of hardware-accelerated real-time data processing for virtualized RAN components
    • Extension of “DevOps” principles to “seamless” software upgrades

Role in project

AP9 plays a central role within the Open6GHub through the development of software technology. As such it is providing the functional basis and the interaction with the hardware for the other work packages more related to network management and optimization as AP4, AP7 and AP8 and the integration with the radio from AP6.  

Results and achievements

  • A basic feasibility study on the possibility of integrating additional software engineering concepts into a new software architecture
  • Definition of a set of internal network services: scaling, migration, upgrades, etc.
  • initial concept design for a complete software network

Resources

This work package concentrates on the development of the software networks aiming to dynamically distribute the system on common, general purpose infrastructures. As such, independence of hardware is a goal. However, specific integrations with acceleration libraries and technologies are considered, to achieve the expected network capacity, especially in the RAN central units and core network data plane. 

Contacts

Prof. Dr. Thomas Magedanz

Head

Dr.-Ing. Marius-Iulian Corici

Deputy Head

Prof. Marina Petrova

Domain Expert

Navid Keshtiarast

Domain Expert

Prof. Dr. Eckhard Grass

Domain Expert

Dr. Lara Wimmer

Domain Expert

Prof. Hans Schotten

Domain Expert

Daniel Lindenschmitt

Domain Expert

Michael Gundall

Domain Expert

Prof. Dr. Martina Zitterbart

Domain Expert

Priv.-Doz. Dr. Roland Bless

Domain Expert

Prof. Dr.-Ing. Matthias Hollick

Domain Expert