ABSTRACT
Alert aggregation is an important subtask of intrusion detection. The goal is to identify and to cluster different
alerts produced by low-level intrusion detection systems, firewalls, etc. belonging to a specific attack instance
which has been initiated by an attacker at a certain point in time. Thus, meta-alerts can be generated for the
clusters that contain all the relevant information whereas the amount of data (i.e., alerts) can be reduced
substantially. Meta-alerts may then be the basis for reporting to security experts or for communication within a
distributed intrusion detection system. We propose a novel technique for online alert aggregation which is based
on a dynamic, probabilistic model of the current attack situation. Basically, it can be regarded as a data stream
version of a maximum likelihood approach for the estimation of the model parameters. With three benchmark
data sets, we demonstrate that it is possible to achieve reduction rates of up to 99.96 percent while the number of
missing meta-alerts is extremely low. In addition, meta-alerts are generated with a delay of typically only a few
seconds after observing the first alert belonging to a new attack instance. Two types of intrusions are detected in
this work: Firstly a spam attack is detected based on the blacklisted IP addresses from Stop Forum Spam and
secondly packet level intrusion is detected based on KDDcup data. A packet sniffer is designed which keeps
sniffing and extracting all the packets that are exchanged over internet interface. The packets are filtered and the
headers are extracted. The headers are further subdivided into TCP, IP and UDP headers. ICMP packets are then
separated. The data is matched with the database intrusion entries using fast string matching techniques and
possible attack entries are marked with different color codes. An attack signature may be visible in any header of
the same packet. In such cases, the alerts are aggregated and a single alert is generated. A signature can be
mutated to multiple packets with similar signature. Such alerts are also combined to a single alert such that the
amount of alert being generated is controlled and that only the signature of the attack is available with the
attacker. Results shows that, adaptation of this technique can not only detect all the signatures with .02% false
acceptance rate and .06% false rejection rate but at the same time can keep the total number of alerts down below
25% of the overall alerts being generated.
Keywords: - Alert aggregation, attacks, meta-alerts, packets, intrusion detection.