6G Systems are expected to support a variety of services over a common infrastructure that is efficiently shared through slicing. Novel Quality of Experience (QoE) architectural models and resource assignment schemes are needed that are able a) to differentiate Service Data Flows (SDFs) originating from the same or multiple user equipment (UE), b) react to changes of the underlying physical infrastructure and c) scale with the number of connected devices. Currently, this functionality is provided by centralized management and network orchestration platforms suffering scalability issues. Thus, future systems are expected to operate in a distributed manner allowing applications to directly intervene in the relevant control processes to guarantee the required QoE. In response to this, the present study focuses on the development of a novel flow assignment scheme supporting applications running at the UEs. The scheme is based on Evolutionary Game Theory (EGT) and allows the application functions (AFs) responsible for the control of the UE applications to take traffic routing and steering decisions in a fully distributed way minimizing charging costs and maximizing their QoE. The charging function of the EGT model relies on profiling data extracted from an open-source 5G platform deployed in a practically implemented testbed environment.

One of the fundamental challenges that needs to be addressed in 5G and beyond systems relates with delivery of demanding uninterrupted services to mobile users. In this context, this paper provides an overview of relevant architectural options taking into consideration user mobility requirements and proposes suitable resource allocation schemes to enable optimal service provisioning under mobility considerations. The proposed schemes are evaluated through a purposely developed multistage optimization framework. This evaluation framework facilitates optimal placement of network and compute resources minimizing the associated infrastructure operational costs. Our results have been produced using real user mobility statistics as well as lab based profiling measurements of the 5G infrastructure and provide a detailed trade-off analysis between latency and infrastructure related costs indicating optimal operational points for different scenarios.