Cost-efficient Placement and Scaling of 5G Core Network and MEC-enabled Application VNFs

The adoption of the 5G technology is becoming a must for the mobile network operators (MNOs) in order to remain competitive in the market and efficiently cope with the increase in the mobile data traffic while providing support for a number of futuristic use cases and services. Given the unprecedented demand of mobile data traffic and its variation, it is of paramount importance to dynamically scale the 5G core network functions as well as the applications, both of which are expected to be deployed as virtual network functions (VNFs), in order to avoid over/under-utilization of these VNFs. In this work, we study the problem of a joint user association, service functions chain (SFC) placement, and VNF scaling with a particular emphasis on analyzing the trade-offs between the VNF scaling strategies. Specifically, we compare vertical, horizontal, and hybrid VNF scaling strategies by formulating an integer linear programming (ILP) problem that aims at minimizing the service provisioning cost for the MNOs, while satisfying users' data rate requirements. The users' service requests are represented as SFCs composed of end-to-end mobile network components (e.g., gNBs, 5G core network VNFs, and application VNFs). Finally, we devise a heuristic algorithm to tackle the scalability issue of the ILP-based approach.