Simulation Study of Routing Mechanism in the Computing-aware Network

Yuhan Zhao,Zheng Chong,Xueying Han,Zongpeng Du,Ke Yu,Xiaohong Huang
DOI: https://doi.org/10.1145/3510513.3510534
2021-01-01
Abstract:Traditional routing algorithms select the path based on the cost, hop, or other network-related metrics. However, with cloud computing and edge computing development, it is essential to develop a computing-aware routing mechanism to distribute the task to the computing nodes with better capabilities. In order to achieve the combined scheduling of computing resources and network resources, Computing-aware Network (CAN) is proposed. The solution aims to comprehensively consider the computing resources and network resources existing when routing and forwarding user requests, which is essential to improve the current situation of uneven utilization of edge computing resources. As many theories and structures are proposed, the requirements for experimental verification of CAN are gradually increasing. However, there is still a lack of experimental simulation prototypes available for study. In this work, motivated by the gap, we propose a simulation prototype based on the idea of the CAN, where routing nodes can incorporate both the computing status and network status into routing and forward packets to an edge server according to the computing demand service type. We verified the effectiveness of the CAN prototype from simulation experiments on a cross-domain topology. From the results, we can conclude that the utilization rate of computing resources has been effectively improved compared to not considering computing status.
What problem does this paper attempt to address?