Research On Dendritic Cell Biology Study Based Cloud Computing Intrusion Detection Technology

Intrusion detection is an important network security technology,which has played a vital role in both dynamic monitoring and network traffic analysis.In recent years,biology-heuristic artificial immune system becomes popular in conventional intrusion detection field,while the dendritic cell intrusion detection model based on danger theory also converting to be a top star with its lightweight detection computation,strong anomaly detection capabilities,and less training costs,etc.It is cloud computing that gradually being recognised by the industry and the public in the process of handling problems related to network data also accompanied with unprecedented growth in the size of large-scale intrusion data,which makes cloud security areas facing enormous challenges.How to solve the intrusion detection in the background under a cloud environment,we need researchers to find innovative thinking ways on the basis of traditional network intrusion detection technology.This paper discusses the cloud environment intrusion detection problems when facing challenges,and proposes a cloud intrusion detection architecture named CIDaa S.The architecture core idea is "Intrusion Detection as a Service".In the experimental results of the response time,we have verified the feasibility of the model deploying in cloud environments.A core research object of our study is the cloud environment overall performance log data,which is derived from the Open Stack Iaa S.Related data dimensions directly reflect the effective attributes when "virtualization",which represents the core technology of cloud,facing the intrusion attacks.We also focus on detailed log data analysis of data preprocessing,and use it in our intrusion detection performance test.Besides,this paper focuses on dendritic cell intrusion detection model.We review the principle of the model,the advantages and limitations of the model.To tackle the limitations of real-time models and multi-model binding limitations,we propose a real-time dendrites twin support vector machine called RDC-TWSVM.In the experiment of traditional platform,we test the training process based on KDD,after that,we put it in the test process with cloud environment overall performance log data,while the empirical results verify the validity of the model and the advanced quality of the intrusion detection performance.Finally,we finished the migration of our proposed intrusion detection model based on Spark platform.Then we build our Spark experiment platform,complete the RDC-TWSVM transplantation,and test the time overhead as well as intrusion detection performance.After the comparison with traditional platform,we verify that the training speed and test speed of the transplantation model have a 20-fold increase,and with a considerable comparative detection performance.The experimental results show that the cloud platform based on Spark could handle the cloud intrusion detection problems in an efficient and practical way.