Navigation

Matthias Frei, M. Sc.

Department of Computer Science
Chair of Computer Science 7 (Computer Networks and Communication Systems)

Room: Room 06.135
Martensstr. 3
91058 Erlangen

More Information

No publications found.

  • Mobility in Multi-Access Edge Computing (MEC) environments

    (Non-FAU Project)

    Term: 2019-10-01 - 2023-09-30

    In the future, data exchange will no more be between a cloud/datacenter-based server and a remote client only. Communication between clients need to be established directly due to aims like immersive applications, autonomous driving or platooning. 5G and future networks are going to follow a data-centric paradigm, respecting the increasing relevancy of direct communication between User Equipment (UEs). On top of that, computing and information resources will no longer be provided by cloud servers only.

    Multi-access edge computing (MEC) ist targeted by current research activities, where resources are located on distributed edge servers. MEC instances may be located close to base stations, to enable applications with special constraints like low-latency, high bandwidth or privacy concerns directly near a UE.

    This prospective concept will be adapted to services that can run literally everywhere – not following a strict hierarchical deployment concept anymore. In addition to a cloud instance, a service can be provided on an edge instance nearby, e.g. a basestation, a traffic control system or even a close-by UE. For energy saving or redundancy purposes it is desirable for an MEC orchestrator to dynamically reallocate edge resources between computing nodes. Next to that, also orchestration decisions and the movement of UEs can lead to frequent changes of the topology. This consequently leads to a mesh-like network setup where communication links are designed and set up based on the data transmitted and not strictly following a pure predefined network design.

    Network Function Virtualization (NFV) and the structural changes applied with the 5G evolution provide more system flexibility: the programmable data plane, part of an User Plane Function (UPF) in the 5G core, does enable the implementation of new ways of data-driven packet handling and forwarding.

    In today’s networks, services are usually deployed as server services located regarding a given network topology in a hierarchical architecture. The limitation in mobility scenarios of current approaches is most-likely the connection-oriented host-to-host communication. Plus, the ability of remote/in-network readdressing of client requests or responses for seamless connectivity, e.g. when the target edge node changes due to an outage, for energy saving reasons or on a handover to another MEC zone, is not natively existent.

    The research project targets visionary communication protocols and setups with focus on mobility scenarios for V2X applications in mesh-like MEC environments. Hereby, realization of Quality of Service (QoS) over network borders and independent network control planes is an essential aspect for the named use cases. Approaches of modern addressing methods and network protocols like Information-Centric Networking (ICN), Locator/Identifier Separation Protocol (LISP) or IPv6 are also considered as relevant.  Summing up, the research project deals with the challenges of network communication in frequently changing, non-hierarchical Service-oriented Architectures (SOAs).