Traffic Prediction Driven Slicing and Reconstruction of 5G Transport Network
Network slicing is one of the key features in 5G mobile communications, and is enabled to support various services with different quality of service (QoS) requirements. However, the high dynamics of traffic make it difficult to fully use the reserved bandwidth of a sliced network. To reduce the operational cost, it is necessary to adjust the slices according to the traffic requirements, to achieve higher bandwidth utilization ratio. This feature will require the 5G slice to be created or adjusted in real-time mode. In 5G transport network, there are different kinds of telecom equipment, which have different control interfaces and protocols. It may need more time for 5G transport network for slice creation and adjustment, so that the slice operation is usually delayed after the traffic change. To overcome this issue, this talk will propose a traffic prediction driven 5G transport slicing scheme. In this scheme, a deep-learning approach is adopted to learn from the previous short-term traffic data, and provides a timely traffic prediction in a sliding window. The prediction change will trigger the controller to re-calculate the underlying network resource allocation and to send the corresponding commands to the transport equipment. Based on the hieratical controller structure, the experiments showed that the mean utilization efficiency was improved obviously.
Networks; network optimization; network slicing; traffic prediction.
Rentao Gu, Full Professor of Beijing Key Laboratory of Network System Architecture and Convergence in Beijing University of Posts and Telecommunications (BUPT), China. He received the Ph.D. degree and Bachelor degree from the Beijing University of Posts and Telecommunications (BUPT) in China, in 2010 and 2005 respectively. From 2008-2009, he studied in Networks and Learning Lab in Georgia Institute of Technology, USA, as a visiting scholar. Since 2010, he joined BUPT as a faculty member. After that, he has been involved more than 7 National Projects related to the optical access networks and software defined optical networks. He has published over 100 peer-reviewed journals (such as IEEE ComMag, IEEE/OSA JOCN, VEHCOM(Elsevier), OFT(Elsevier), OSN(Elsevier), Europhysics Letters, Physica A) and conference papers (OFC, ECOC etc.), and hold 25 granted China and US patents. He has been a reviewer of many IEEE/Elsevier/IET journals and conferences, and involved in the various conference organizing as publicity co-chair, TPC co-chair, and OC/TPC members. He is a senior member of IEEE, China Institute of Communications (CIC) and Chinese Institute of Electronics (CIE), and member of CIE Senior Member Evaluation Board; and he was Young Professional Standing Committee Member of IEEE Communication Society (2016-2017).