Service and network function placement in the edge-cloud continuum

The virtualization and containerization technologies have paved the way towards a flexible and fully programmable service provisioning chain over mobile telecommunication infrastructures. On top of those technologies the network slicing concept has emerged; a concept that enables the installation and migration of service and network functions on a compute fabric that spans from the central cloud to distributed far-edge hosts. In this context, the efficient placement of service and network functions in the hosts of the continuum is an open and challenging problem. We formulate this problem as a scheduling process and we analyze its optimal performance through the integer linear programming toolkit. The study revealed interesting insights regarding the impact of sequence constraints (i.e., some functions should be allocated according to a specific order) and host-continuity constraints (i.e., some functions should be allocated in the same cloud host with others) to the overall efficiency of the scheduling.