Routing Policy for Balanced Management of Slices Using Flexible Ethernet.

Flexible Ethernet (FlexE) technology meets the business needs of multi-granularity bandwidth, low latency, high reliability. With the increasing demand for small-granularity businesses, the equipment management burden and computing capacity limitations caused by the huge increase of small-granularity slices should be considered. We need to achieve a balanced allocation of managed resources while ensuring deterministic latency and high reliability of the network. Therefore, this paper presents a multi-granularity end-to-end routing policy for balanced management of slices. We model this transport scenario of multi-granularity slices and abstract the balanced management routing of the slices. Then, using the method of Reinforced Learning (RL), an effective algorithm is derived to optimize capacity utilization and management balance coefficient by balancing the management resources of slices. The simulation results show that the algorithm can achieve a good balance effect, get a more average resource management scheme, reduce the error loss rate of slices during many business transports, and effectively improve the network performance.