An Innovative Signature Detection System for Polymorphic and Monomorphic Internet Worms Detection and Containment
Most current anti-worm systems and intrusion-detection systems use signature-based technology instead of anomaly-based technology. Signature-based technology can only detect known attacks with identified signatures. Existing anti-worm systems cannot detect unknown Internet scanning worms automatical...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | eng eng |
| Published: |
2012
|
| Subjects: | |
| Online Access: | https://etd.uum.edu.my/3353/1/MOHAMMAD_M._RASHEED.pdf https://etd.uum.edu.my/3353/3/MOHAMMAD_M._RASHEED.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Most current anti-worm systems and intrusion-detection systems use signature-based technology instead of anomaly-based technology. Signature-based technology can only detect known attacks with identified signatures. Existing anti-worm systems cannot detect unknown Internet scanning worms automatically because these systems do not depend upon worm behaviour but upon the worm’s signature. Most detection algorithms used in current detection systems target only monomorphic worm payloads and offer no defence against polymorphic worms, which changes the payload dynamically. Anomaly detection systems can detect unknown worms but usually suffer from a high false alarm rate. Detecting unknown worms is challenging, and the worm defence must be automated because worms spread quickly and can flood the Internet in a short time. This research proposes an accurate, robust and fast technique to detect and contain Internet worms (monomorphic and polymorphic). The detection technique uses specific failure connection statuses on specific protocols such as UDP, TCP, ICMP, TCP slow scanning and stealth scanning as characteristics of the worms. Whereas the containment utilizes flags and labels of the segment header and the source and destination ports to generate the traffic signature of the worms. Experiments using eight different worms (monomorphic and polymorphic) in a testbed environment were conducted to verify the performance of the proposed technique. The experiment results showed that the proposed technique could detect stealth scanning up to 30 times faster than the technique proposed by another researcher and had no false-positive alarms for all scanning detection cases. The experiments showed the proposed technique was capable of containing the worm because of the traffic signature’s uniqueness. |
|---|