120 likes | 307 Views
Intrusion Detection for Web Services. Yang Sha ys2280@columbia.edu. Intrusion detections. Host-based IDS (HIDS) Scan logs on local system Network-based IDS (NIDS) Analyze network packets Impossible to check all packets Misuse detection IDS Rule-based IDS Anomaly detection IDS
E N D
Intrusion Detection for Web Services Yang Sha ys2280@columbia.edu
Intrusion detections • Host-based IDS (HIDS) • Scan logs on local system • Network-based IDS (NIDS) • Analyze network packets • Impossible to check all packets • Misuse detection IDS • Rule-based IDS • Anomaly detection IDS • Learn the patterns of user and malicious behaviors
Threat models for web services • XML DoS and DDoS • Jumbo payloads, Recursive elements, XML flood • Unauthorized access • Data integrity • XML hijacks • System compromise • Attacker gets the admin permission
XML DoS Example <?xml version="1.0" encoding="UTF-8"?> <!DOCTYPE root [ <!ENTITY ha "Ha !"> <!ENTITY ha2 "&ha; &ha;"> <!ENTITY ha3 "&ha2; &ha2;"> <!ENTITY ha4 "&ha3; &ha3;"> <!ENTITY ha5 "&ha4; &ha4;"> ... <!ENTIY ha128 "&ha127; &ha127;"> ]> <root>&ha128;</root> </xml>
IDS solution to XML DoS and DDoS • Rule-based NIDS (like Snort) • Sample rules for detecting DoS and DDoS
XML Hijack Example • One of the most common attacks
IDS solution to XML hijack • Trying all the possible cryptography algorithms • Very time-consuming
IDS solution to other data integrity issues • Impossible for network based IDS to identify the data integrity • Without private key or session key, IDS can hardly analyze the encrypted XML documents
IDS solution to unauthorized access • Build up the user behavior pattern based on past operations • Compare new behaviors with past patterns, and pop up alert whenever there is a huge mismatch
Conclusion • IDS can only solve part of the problem • Hard to avoid zero-day attack • Future improvement • packet sampling • Real-time classification of user behavior patterns