Design And Research Of Campus Network Security Based On Intrusion Detection Technology

With the proposal of “Cyberpower” concept in our country,the deployment of network equipment is growing year by year.The business volume of e-commerce,online games,document services,and multimedia services have also gradually increased.For defensing the Covid-19,location services have also shown a growth trend.The growth of these internet services has brought great convenience to people's production and life.At the same time,due to the presence of malicious attackers,these network traffic has also brought information security risks and potential information leakage risks to more and more users.This type of network security threats is particularly obvious in campus networks.Due to the continuous improvement of hackers' attack methods and the increasing number of attacks,they have brought great inconvenience and security threats to the school's daily administrative work,teaching work,and students' daily internet access.Therefore,researching and constructing a proper method for detecting abnormal data flow of campus network is an essential part of campus network security,which has great significant to guarantee the security of campus networks.Based on the current network architecture of Changchun University of Science and Technology,this paper proposes a PCA and random forest-based system to identify the abnormal flow of campus networks,which is aim to resolve the low detection accuracy,low detection efficiency problems.The main contents of this paper are included as follows:1.The data collection system of Changchun Institute Science and Technology was built in this paper,the topology of the campus network of Changchun Institute of Science and Technology was designed firstly,and the collected data was marked and restored according their category.In order to save cost and not change the topology of the current networks,this paper uses port mirroring technology to copy and forward the data packets collected from the backbones.The automatically store and analysis system was also designed.To evaluate the data collection system,we use Kali Linux system test the network in random time,experimental results shows that our system could capture the data packet from backbone networks correctly.2.Aim on the over-fitting and low efficiency problem in intrusion detection system,this paper proposes to use principal component analysis(PCA)method to decrease the dimensionality and extract the feature,and we test the accuracy of the PCA with CIC-IDS dataset,the experimental results show that the use of the PCA method improves the detection rate of the intrusion detection system for abnormal data streams and improves the detection efficiency.3.The random forest based abnormal data flow classifier was proposed in this paper,and we verified the accuracy of the algorithm on the CIC-IDS 2017 data set and the campus network data set of Changchun Institute of Science and Technology.The experimental results show that the accuracy random forest algorithm could achieve99.82% and 96.81 in CIC-IDS 2017 database and database collected from Changchun University of Science and Technology,which is better that the state-of-the-art technologies.Experimental results show that the PCA and random forest-based intrusion detection system could effectively detect the abnormal data flow of campus network,and has certain theoretical value and practical application reference value.