Abstract
Network slicing is a key technology that enables support for a wide range of services on a common communication infrastructure. Although end-to-end network slice allocation to multiple tenants has been studied in the existing literature, the user-slice assignment is still a challenging problem due to its combinatorial nature. In this paper, we study the joint problem of user-slice assignment and intra-slice resource pooling to maximize the support for the expected traffic of users, minimize the expected outage of downlinks, and maximize isolation of unpooled reserved resources to serve users. We propose a block coordinate descent algorithm based on successive upper-bound minimization to simultaneously find the user-slice assignment and allocate resources to slices in the backhaul and Radio Access Network (RAN). Simulation results verify the efficiency of our approach against existing work.