Secure Server System

Abstract

Secure Server System(SSS) is server based application which runs on network layer because security of server always faces new potential threats as hackers and viruses advance. It will filter all the incoming data packet and identify known and unknown intrusions, virus or threats. This application is good for server security. Develop a server application and filtering the data packets at network layer and implemented on the server. The purpose of server security is to provide confidentiality, Integrity and Availability (CIA). Protect the privacy and integrity of messages going through untrusted networks. It will also prevent a number of attacks like Denial-of-serves attack, Spear Phishing Attacks, Phishing Attacks, man-in-the- middle attack etc. In secure server system application, we use IDS and IPS through this mechanism application detect the intrusion and prevent those intrusions. Monitoring and traffic, scanning the data packet and ports for suspicions data. This application gives all the information related to the server and network like how much ports are open and how much ports are closed and gives all data packets information like which data packet coming from which source and reached to which destination. One of the major research challenges in this field is the unavailability of a comprehensive network based data set which can reflect modern network traffic scenarios, vast varieties of low footprint intrusions and depth structured information about the network traffic. Today Firewall systems cannot identify modern attack environments and are not able to analyses network packets in depth. Because of these reasons, IDSs are designed to achieve high protection for the cyber security infrastructure. These algorithms are used in the classification the Intrusion Detection Attacks. Here, three classification models such as Random Forest and Multilayer Perceptron are used and compared their performance. Classification of attacks are made and by applying the evaluation criteria the corresponding Specificity, Accuracy, Sensitivity are evaluated to get the respective True Positive, false positive rate for all mentioned algorithms. Sensitivity for MLP 1.0%, Random Forest(RF) 0.999% and Logistic Regression(LR) are 0.998% respectively. Specificity for MLP 1.0%, RF 0.999% and LR are 0.908% respectively. Accuracy for MLP 1.0%, RF 0.999% and LR are 0.994% respectively.