Abstract:For the last years intrusion detection system(IDS) has been proved valuable in protecting computer systems against malicious attacks. It works by actively detecting abnormal activities and counter act,be it reporting to the user or taking certain actions automatically. The traditional IDS,however,is defective in that it deals with the detected abnormal activites only individually,bringing serious problems,classical ones of which include massively false alarms and massively repeated low-level alarms. An even more serious problem of the traditional IDS is that it is not effective,if not uselell,when the many individually detected abnormal activites are correlated components of a multi-step indepth intrusion,which is propably the case in nowadays since intrusions have evolved to be intelligent and distributive. To overcome this defect,this paper proposes a method called MSACA(multi-step attack correlating algorithm). As a prerequisite of the method,a knowledge database is provided. The knowledge database composes of the structures of known multi-step correlating attack,meaning that given certain low-level abnormal activites,through the database,we can get information regarding to what kinds of multi-step correlating attacks the given low-level alarms may be part of as well as to which stage they belong. The proposed method firstly gets the needed information of the low-level alarms from the database,secondly tries to form a big picture to reveal the possible threat of the low-level alarms working as a whole,and finally takes counter measures. The proposed method can be implemented in realtime scenarios. This paper also proves the proposed method's effectiveness by experiments.

Statistical Decision Modeling for IDS Alert Analysis

Use of host's software information to mitigate false positive in signature-based NIDS

Study of an IDS Alert Fusion and Correlation Model

Alert reduction for network intrusion detection

A Novel Data Mining-Based Method for Alert Reduction and Analysis.

A Novel IDS Evaluation Model Based by Bayes Estimation

Using Outlier Detection to Reduce False Positives in Intrusion Detection

Reducing False Positives Based on Time Sequence Analysis

Alert Correlation by A Retrospective Method

An intrusion detection method based on decision tree

Discovering Anomaly on the Basis of Flow Estimation of Alert Feature Distribution

An Entropy-based Method for Attack Detection in Large Scale Network

Processing Intrusion Detection Alerts in Large-scale Network

An Approach to Filter False Positive Alerts Based on RS-SVM Theory

A New Attack Detection in Large Scale Network Based on Entropy.

Alert Correlation Through Results Tracing Back to Reasons

Enhanced Intrusion Detection Systems Performance with UNSW-NB15 Data Analysis

Research on Decision-Making of IDS with Multiple Parameters

An alert correlating algorithm based on multi-step attack path graphs:MSACA

Research on the Method of Reducing False Positives with Periodicity

Alert Correlation Based on Description Logics