BCLB: Blockchain-based Controller Load Balance for Safe and Reliable Resource Optimization.

Hao She, Xiaozhen Zhu, Yongan Guo,Haotong Cao,Sahil Garg,Georges Kaddoum

GLOBECOM (Workshops)(2022)

引用 1|浏览11
暂无评分
摘要
Controller in software-defined network (SDN) can realize the intelligent load balancing of switches. However, the controller is lacking in the ability of global load management and control methods for managing multi-domain controllers. With the benefits of realizing decentralized, traceable, and trusted data sharing, blockchain can provide controllers with information such as inter-domain links and the global view to assist achieving global controller load balancing. We propose one global controller load-balancing method based on blockchain in this paper. Our proposed method adopts in-band telemetry (INT) so as to sense the status of each network domain, and combine the load of the controller itself. This enables to package and upload the chain to form a decentralized data-sharing model. Based on the load and network domain status of the global controller, combined with deep reinforcement learning (DRL) algorithm analysis, a specific solution is obtained, and some edge switches in the network domain of the overloaded controller are selected for migration, thereby achieving global load balancing. The method in this paper can not only integrate global information and reduce migration costs, but also ensure that domain information is not leaked. Simulation results show its convergence and effectiveness.
更多
查看译文
关键词
controller load balance,blockchain-based
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要