Application Layer Proxy Detection, Prevention with Predicted Load Optimization Vijaypal Singh Dhaka Vikas Lamba Anubhav Divyanshu Pathania Computer Science Computer Science Computer Science Computer Science Manipal University Jaipur National University Jaipur National University Jaipur National University Jaipur, India Jaipur, India Jaipur, India Jaipur, India [email protected] vlamba [email protected] [email protected] [email protected] Abstract—In this paper, we have formulated a solution for proxy usages in a network. We all are surrounded with digital signatures around us, we just need to filter those to make our network clean, secure, efficient. Is what we did in this research work. There is also a great need of load optimization at different points in the network. We have proposed a way to use machine learning and neural networks to apply it on a network. Our basic approach to do this work is a time quantum analysis for load prediction over a network and digital signature validation for proxy detection at the application layer. There are many methods to detect network statistics and security at different layers, but we need real-time analysis in the network and prevention measure right before it happens. Keywords— Neural Networks, Digital Signature, Machine Learning, Computer Networks I. INTRODUCTION In the TCP model of computer networks whole communication, we have different layers and each layer has its own functionality. This model is responsible for the overall functioning of a network. There are different forms of attacks, intrusion and proxy in our network. Computer network security is a very important concern in terms of reliable, secure, efficient. Attack in any network strikes through outside the network or it can also strike network through inside. Hacks is a similar situation network attack occurring through outside the network. The proxy is situation in which network intrusion occurs from inside the network. In our research, we have formulated the way to detect Network Intrusion which occur due to many proxy software which is mainly to use to create proxy connection with remote server and used to unauthorized contents over a network. There are many types of proxies like virtual private network, proxy over the browser, proxy software. Also predict network load using machine learning techniques. Proxy have no limits; they are growing day by day. Every day there occurs new proxy server around the world. Network load is also very important factor more efficient functioning of the network. Load prediction is needed is every network as the growing nature of network in current world. Human can’t intervene the hardware in real-time We need a technique to manage the load on each router with predictive analysis and statistics. A. How Proxy Works Proxy provides an interface for the user and the blocked websites by redirecting the content on that server to his own server first and then transmit those packets to user. Proxy Server: A proxy is a remote server which acts as a middle server between the user and the main server. Proxy main work is to provide the blocked content to the user which is unable to access in his own network. So a proxy server routes the packet from requesting server to the user in his network. Proxy Server Detection: One of a solution for detecting the proxy server is to blacklist them in your own network so if any one creates request to remote server with the help of proxy server can easily be detected and thus increasing network reliability.