Abstract : Network slicing is one of the cornerstone features of the fifth generation (5G) networks to enable the multiplexing of virtualised logical networks over the same physical network infrastructure for various vertical business services. Architectures based on network slicing enable network operators to offer end-to-end (E2E) vertical services whilst assuring fulfilling tailored service level agreement (SLA) requirements. Despite the numerous benefits that this concept offers, the need for efficient, scalable and holistic E2E management of network slicing significantly complicates network operators’ network management and imposes significant challenges in the quality of service (QoS). This paper proposes a novel intent-based slice manager framework over a service base 5G architecture that allows flexible definitions to customise network slices, automates all essential management tasks for 5G network slice providers, manages the life cycles of all sorts of network slices, and guarantees their QoS in a unified network slice management framework. The proposed framework is empirically validated in a realistic large-scale 5G multi-tenant infrastructure. The high scalability of the framework is based on the empirical results supporting more than 512 physical machines, 534288 virtual machines and more than two million network interfaces.
Index terms : Automation, end-to-end network management, network slicing, 5G networks.