TITLE:
Container Networking Performance Analysis for Large-Scale User Behavior Simulation
AUTHORS:
Yifang Ji, Guomin Zhang, Shengxu Xie, Xiulei Wang
KEYWORDS:
Linux Container, Networking Mode, Network Performance, User Behavior Simulation
JOURNAL NAME:
Journal of Computer and Communications,
Vol.7 No.10,
October
16,
2019
ABSTRACT: Accurately simulating large-scale user behavior is important to improve the similarity between the cyber range and the real network environment. The Linux Container provides a method to simulate the behavior of large-scale users under the constraints of limited physical resources. In a container-based virtualization environment, container networking is an important component. To evaluate the impact of different networking methods between the containers on the simulation performance, the typical container networking methods such as none, bridge, macvlan were analyzed, and the performance of different networking methods was evaluated according to the throughput and latency metrics. The experiments show that under the same physical resource constraints, the macvlan networking method has the best network performance, while the bridge method has the worst performance. This result provides a reference for selecting the appropriate networking method in the user behavior simulation process.