Colloquium lecture: 28 February 2023, Parham Gandomkar (A. Brummer)

Symbolic picture for the article. The link opens the image in a large view.
Bild der Präsentationsfläche

Real-Time Optimization of MEC-enabled Communication Networks

The MEC demand is driven by service creators, whose services have strong network requirements. For example, in connected driving scenarios, where (external) sensor data is distributed and shared between vehicles and other road users, the communication needs to be reliable and very fast. Also in other service domains like audiovisual communication (holographic calls, AR/VR) or Artificial Intelligence(AI)-based applications, MEC-support can enable new possibilities and use cases. However, in future network topologies, numerous MEC resources might be available to select. A suitable MEC resource must be determined for each client – with regard to the service’s communication requirements and the service provider’s resource allocation strategy. Hereby, one of the most important key properties must be taken into account: A MEC service has strict communication constraints compared to a cloud service, such as maximum latency, maximum tolerable jitter, minimum availability rates or privacy requirements. When optimizing the client-to-MEC associations as well as the service deployment locations (e.g., 1, 2, 3, 4), the current network conditions should be considered. Therefore, in this master thesis, an approach that optimizes selected goals globally and applies new configuration in real-time should be implemented, validated and tested. The current network situation should be determined by measurements of selected KPIs on various network infrastructure nodes and be are reported to a network-global database. The information from the database should then be used to run the optimizations. Finally, the resulting associations respectively the new network configuration should be applied and tested. For realistic testing scenarios, key requirements of real applications should be determined from current literature and a scenario that enables representative comparisons should be defined and implemented. With service KPIs defined, an application communicating with these characteristics should be implemented as an OMNet++ app and be used to evaluate the current network configuration.

 

Time: 10:15 am

Place: room 04.137, Martensstr. 3, Erlangen

or

Zoom:
https://fau.zoom.us/j/63174613617?pwd=SFBCVE4vckZsekkxSEQxSXBqVkZwdz09

Meeting-ID: 631 7461 3617
Kenncode: 044678