Business Demand-Oriented Intelligent Orchestration of Network Slices Based on Core-Edge Collaboration

5G enables many industries, and each industry's application requirements vary greatly. Today's “one-size-fits-all” network architecture approach no longer meets the needs at the same time. The introduction of network slicing brings great flexibility to the network, so that the network can be customized, deployed and dynamically guaranteed. However, how to orchestrate the functions of network slices according to the needs of business scenarios is an important challenge for slice operation. In order to solve the problem that the existing methods can hardly distinguish the differential requirements of different slices for delay, bandwidth and node load balancing, this paper proposes an exclusive orchestration optimization goal to match the business requirements, and establishes a business demand-oriented network slice orchestration problem model. Then, aiming at the typical slices, a business-oriented slicing algorithm based on DQN (BOSAD) is proposed. In this algorithm, we propose the strategy of cooperating the core data center and the edge data center, which can effectively save bandwidth and reduce network delay. The experimental results show that the proposed BOSAD saves the consumption of bandwidth resources, reduces the average slice delay and optimizes node load balancing.