A Novel Anti-DDoS Microservice Load Balancing Scheme Based on Improved GBDT Algorithm

DDoS is a common attack mode that affects Internet security. In order to deal with a large number of DDoS attacks, it is commonly used to deploy anti-DDoS programs in the form of microservices. However, there is no way to achieve a good microservice load balancing effect. This paper proposes an anti-DDoS microservice load balancing scheme based on improved GBDT algorithm. By taking advantage of the correlation between server response time and server CPU utilization, memory utilization, disk utilization and task type, a machine learning regression model is established based on improved GBDT algorithm to predict the response time when a new task arrives. The anti-DDoS task is distributed to the servers in the microservice cluster with the shortest predicted response time, so that the server load in the cluster is relatively balanced. The experimental results show that the scheme can achieve better load balancing effect and has the value of deployment and popularization.