A Comprehensive Survey of Intrusion Detection Systems

Abstract

Alongside with digital signatures and Cryptographic protocols, Intrusion Detection Systems (IDS) are judged to be the final contour of protection to protect a system. But the major difficulty with today’s mainly admired IDSs (Intrusion Detection System) is the invention of massive quantity of false positive (FP) alerts alongside with the true positive (TP) alerts, which is an awkward assignment for the operator to examine to arrange the proper responses. So, there is an immense requirement to discover this area of study and to discover a reasonable solution. A main disadvantage of Intrusion Detection Systems (IDSs), despite of their detection method, is the vast number of alerts they produce on a daily basis that can effortlessly exhaust security supervisors. This constraint has guide researchers in the IDS society to not only extend better detection algorithms and signature tuning methods, but to also focus on determining a variety of relations between individual alerts, formally known as alert correlation. There are a variety of approaches of intrusion detection, such as Pattern Matching, Machine Learning, Data Mining, and Measure Based Methods. This paper aims towards the proper survey of IDS so that researchers can make use of it and find the new techniques towards intrusions.

Keywords: Intrusion Detection System, False positive alert, KDD Cup99, Anomaly detection, misuse detection, Machine Learning.