Machine Learning Based Traffic-Aware Service Reconfiguration in Metropolitan Area Networks

Zhu, Qingcheng; Wang, Yuqiao; Meng, Xiangyu; Zhao, Yongli; Yu, Xiaosong; Zhang, Jie

doi:10.1109/icicas48597.2019.00143

Cited by 2 publications

(2 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Also in the Metro-Haul project, machine-learning (ML) has been used in the allocation resources that serve different services with dynamic traffic [108]. On the other hand, with respect to traffic engineering solutions, recurrent neural network (RNN) have been used to forecast traffic in different areas of the metropolitan network [58], while backpropagation techniques have been used to predict trends in changes in traffic load [107]. In this way, new opportunities to explore AI solutions arise in the field of unbundled networks and communication networks for 5G, taking advantage of centralized control and optimizing operations on these networks [15].…”

Section: Future Trendsmentioning

confidence: 99%

See 1 more Smart Citation

Metropolitan Optical Networks: A Survey on New Architectures and Future Trends

Sousa¹,

Drummond²

2022

Preprint

View full text Add to dashboard Cite

Metropolitan optical networks are undergoing major transformations to continue being able to provide services that meet the requirements of the applications of the future. The arrival of the 5G will expand the possibilities for offering IoT applications, autonomous vehicles, and smart cities services while imposing strong pressure on the physical infrastructure currently implemented, as well as on static traffic engineering techniques that do not respond in an agile way to the dynamic and heterogeneous nature of the upcoming traffic patterns. In order to guarantee the strictest quality of service and quality of experience requirements for users, as well as meeting the providers' objectives of maintaining an acceptable trade-off between cost and performance, new architectures for metropolitan optical networks have been proposed in the literature, with a growing interest starting from 2017. However, due to the proliferation of a dozen of new architectures in recent years, many questions need to be investigated regarding the planning, implementation, and management of these architectures, before they could be considered for practical application. This work presents a comprehensive survey of the new proposed architectures for metropolitan optical networks. Firstly, the main data transmission systems, equipment involved, and the structural organization of the new metro ecosystems are discussed. The already established and the novel architectures are presented, highlighting its characteristics and application, and comparative analysis among these architectures is carried out identifying the future technological trends. Finally, outstanding research questions are drawn to help direct future research on the field. Among the conclusions of this survey, the application of flexible spectrum allocation technologies seems to be the right, and necessary, evolution path for metropolitan optical network architectures.

show abstract

Section: Future Trendsmentioning

confidence: 99%

“…The BPANN service reconfiguration strategy based on traffic-aware prediction model with machine learning proposed in [107] can be applied to design other logical topologies without increasing CapEx/OpEx in order to exploit legacy infrastructure, while solving the unbalanced network resource in metropolitan networks.…”

Section: Hns/wdmmentioning

confidence: 99%