Network Security: Approach Based on Network Traffic Prediction

Abstract

Considering the network security aspect, one of the best way of preventing network infrastructure against anomalous activities is to monitor its traffic for suspicious activities. The reliable resource to accomplish this task is past network flow data, which can be analyzed to detect congestions, attacks or anomalies to ensure effective QoS of network infrastructure. Network traffic prediction involves analysis of past network flow data by capturing-storing data, preprocessing data, analyzing it based on various parameters & forming behavior patterns for various nodes in network. Once the patterns are observed for different nodes in network, their future communication can be predicted. Upon prediction of anomalous behavior, the preventive action will be initiated without wasting much of a time. Thus reducing the MTTR (mean time to respond) is the outline of our paper. The importance of network traffic data, traffic prediction methods and literatures available on topic are studied in this paper.