Trapping resource orchestration scheme based on electric power environment
Tianfu Xu, Tianzhe Li, Qun He, Zhiyuan Luo, Naqin Zhou, Haoqin Chen, Rixuan Qiu · 2024
In the field of network security for power systems, honeypot technology is a crucial means of protecting critical infrastructure from attacks. However, with the increasing number and variety of honeypots, traditional manual management methods have become complex and inefficient. Deploying honeypots using Kubernetes (K8s) has emerged as a new trend. Yet, the native K8s scheduler has limitations in handling the resource load fluctuations caused by honeypot attacks, which can lead to resource wastage and performance instability, affecting the effectiveness and stability of honeypots. To address this issue, this paper proposes a resource orchestration scheme for honeypots in power environments: optimizing the scheduling strategy of honeypots in K8s clusters using a Honeypot-Oriented Kubernetes Scheduler for Dispersing Peak Loads (HKSDPL) by predicting the load fluctuations of honeypots. By establishing a predictive model to capture the load fluctuations, analyzing the resource consumption patterns of honeypots, and adjusting the scheduling decisions accordingly, it avoids the risks of resource over-limit and resource contention. Experimental results show that HKSDPL can increase the number of honeypot deployments by 58.06% without exceeding resource limits, significantly reducing resource contention during high-load periods, and enhancing the stability and security of the power system.