An Efficient Online Heuristic for Mobile Network Slice Embedding
2020 23rd Conference on Innovation in Clouds, Internet and Networks and Workshops (ICIN)
The fifth generation (5G) of mobile networks facilitates the management of various new use cases like autonomous driving, eHealth and the Internet of Things (IoT) along with dense sensor networks. Those use cases induce diverse requirements on the mobile network infrastructure. Network Slicing is one of the key features in 5G to address this issue. Network Slicing organizes the applications in network slices, whereby they are isolated, logical end-to-end networks for different use cases,
... t use cases, sharing a common physical network infrastructure. They can be set up, changed and deactivated dynamically and in realtime. Therefore, efficient Network Slice Embedding algorithms are required for the assignment and allocation of network slice resources. In this paper, a novel, efficient approach for such an algorithm is proposed. It is a heuristic which meets the specific requirements of Network Slice Embedding, based on the algorithm provided by Cheng et al. In addition to their topology-aware node ranking algorithm, three alternative ranking techniques for the node mapping are introduced. The Network Slice Embedding heuristic and its ranking algorithms are evaluated on three generic blueprints of network slices and multiple mobile network sizes. The results show that the proposed heuristic is suitable for real-time embedding as it is robust, flexible and scalable and achieves high acceptance ratios.