Research On Reinforcement Learning Based DDoS Mitigation Method In IoV Under Mec

As a significant part of intelligent transportation system,internet of vehicles(IoV)has been seen as a promising technology in services,such as road safety,collision avoidance and information interactions.Due to the demand for real-time data processing of IoV,many studies utilize the technology of mobile edge computing(MEC)to the building of IoV.However,the distributed denial of services(DDoS)attacks become a serious problem in IoV under MEC.Although numbers of studies have been done on DDoS d etection in common wired or wireless networks,they cannot satisfy the high dynamic requirement and cannot cope with the complex and diverse DDoS attacks in IoV.Fortunately,the data traffic flows in IoV exist potential and predictable space-time regularities.By employing the regularities,we propose a feature adaption reinforcement learning approach based on the space-time flow regularities in IoV for DDoS mitigation,named FAST,to achieve a better DDoS mitigation result..In FAST,we elaborately design a combinational action space,and a reward function based on Kalman filter method and historical data traffic flows,which can make FAST to distinguish DDoS attacks more quickly and accurately.Then through combining Q-learning and DDQN,FAST method can select features and disconnect DDoS attacks adaptively according to the changes of the environment.In addition,although some intelligent reinforcement learning based methods have been introduced to mitigate DDoS attacks,there are still many constraints in the training process,such as the long training time and dependence on large labeled data.Under a 5G ultra-dense network scenario,increasing numbers of micro MEC base stations will be deployed on IoV at any time.In fact,for a new base station without any background knowledge,the agent(i.e.,MEC station)will spend a lot time learning DDoS mitigation strategies for the current environment from scratch.Therefore,how to accelerate the learning speed for these newly added base stations is a challenge.In this paper,by utilizing the knowledge obtained by adjacent similar base stations,we design a Transfer DDQN method based on base station similarity to speed up the training of DDQN agent for a newly added base station.This enables the new base stations to obtain DDoS mitigation policies quickly.Meanwhile,this paper also proposes a method of measuring the similarity of base stations in order to reasonably select the object s to be transferred.In the experiment,we evaluate the performance of the algorithms proposed in this article based on an open source tool-ddosflowgen and Shenzhen taxicab dataset.The experimental results show that FAST method can adaptively select features to defend against multiple types of DDoS attacks without labeled data set.And with the training of the DDQN agent,the algorithm can identify DDoS attacks with high accuracy in a short time under the dynamic environment of IoV.In addition,the DDoS mitigation method based on Transfer DDQN can well obtain the strategies from similar base stations,reduce the training time of new base stations and accelerate the convergence speed of reinforcement learning agent,in order to quickly obtain a DDoS attack mitigation strategy adapted to the current environment of IoV.