CYBER THREAT ANALYSIS A KEY ENABLING TECHNOLOGY FOR

CYBER THREAT ANALYSIS – A KEY ENABLING TECHNOLOGY FOR THE OBJECTIVE FORCE (A CASE STUDY IN NETWORK INTRUSION DETECTION) Vipin Kumar Army High Performance Computing Research Center Department of Computer Science University of Minnesota http: //www. cs. umn. edu/~kumar Authors: Aleksandar Lazarevic, Paul Dokas, Levent Ertoz, Vipin Kumar, Jaideep Srivastava, Pang-Ning Tan Research supported by AHPCRC/ARL

Cyber Threat Analysis w As the cost of information processing and Internet accessibility falls, military organizations are becoming increasingly vulnerable to potential cyber threats such as network intrusions w. There is an increasing awareness around the world that cyber strategies can be a major force multiplier and equalizer

Intrusions in Military and Government Organizations w Intrusions are actions that attempt to bypass security mechanisms of computer systems. They are caused by: w Attackers accessing the system from Internet w Insider attackers - authorized users attempting to gain and misuse non-authorized privileges w Typical intrusion scenario Computer Network Scanning activity Attacker Machine with vulnerability

Intrusions in Military and Government Organizations w Intrusions are actions that attempt to bypass security mechanisms of computer systems. They are caused by: w Attackers accessing the system from Internet w Insider attackers - authorized users attempting to gain and misuse non-authorized privileges w Typical intrusion scenario Computer Network Attacker Compromised Machine

Why We Need Intrusion Detection Systems in Military and Government Organizations w Security mechanisms always have inevitable vulnerabilities w Current firewalls are not sufficient to ensure security in military networks w “Security holes” caused by allowances made to users/programmers/administrators w Insider attacks w Multiple levels of data confidentiality needs multi-layer protection in firewalls

Intrusion Detection w Intrusion Detection System w combination of software and hardware that attempts to perform intrusion detection w raises the alarm when possible intrusion happens w Traditional intrusion detection system IDS tools (e. g. SNORT) are based on signatures of known attacks w Limitations w Signature database has to be manually revised for each new type of discovered intrusion www. snort. org w They cannot detect emerging cyber threats w Substantial latency in deployment of newly created signatures across the computer system

Data Mining for Intrusion Detection § Misuse detection § Predictive models are built from labeled data sets (instances are labeled as “normal” or “intrusive”) § These models can be more sophisticated and precise than manually created signatures § Unable to detect attacks whose instances have not yet been observed § Anomaly detection § Identifies anomalies as deviations from “normal” behavior § Potential for high false alarm rate - previously unseen (yet legitimate) system behaviors may also be recognized as anomalies § Recent research § Stolfo, Lee, et al; Barbara, Jajodia, et al; James; Lippman et al; Bridges et al; etc.

Key Technical Challenges 1. Large data size 1. Millions of network connections are common for commercial network sites, … 2. High dimensionality 1. Hundreds of dimensions are possible 3. Temporal nature of the data 1. Data points close in time - highly correlated 4. Skewed class distribution “Mining needle in a haystack. So much hay and so little time” 1. Interesting events are very rare looking for the “needle in a haystack” 5. Data Preprocessing 1. Converting network traffic into data 6. High Performance Computing (HPC) is critical for online analysis and scalability to very large data sets

The MINDS Project w. MINDS – MINnesota INtrusion Detection System w. Learning from Rare Class – Building rare class prediction models w. Anomaly/outlier detection w. Characterization of attacks using association pattern analysis Rules Discovered: {Milk} --> {Coke} {Diaper, Milk} --> {Beer} M I N D S

M I N D S MINDS - Anomaly Detection Detect novel attacks/intrusions by identifying them as deviations from “normal”, i. e. anomalous behavior w w Identify normal behavior Construct useful set of features Define similarity function Use outlier detection algorithm § Nearest neighbor approach § Density based schemes § Unsupervised Support Vector Machines (SVM)

Experimental Evaluation w Publicly available data set w DARPA 1998 Intrusion Detection Evaluation Data Set § prepared and managed by MIT Lincoln Lab § includes a wide variety of intrusions simulated in a military network environment w Real network data from w University of Minnesota Anomaly detection is applied Open source signaturebased network IDS u 4 times a day network u 10 minutes time window www. snort. org 10 minutes cycle 2 millions connections net-flow data using CISCO routers M I N D S Anomaly scores MINDS Data preprocessing anomaly detection … … Association pattern analysis

Feature construction Three groups of features w Basic features of individual TCP connections § source & destination IP/port, protocol, number of bytes, duration, number of packets (used in SNORT only in stream builder module) w Time based features § For the same source (destination) IP address, number of unique destination (source) IP addresses inside the network in last T seconds § Number of connections from source (destination) IP to the same destination (source) port in last T seconds w Connection based features § For the same source (destination) IP address, number of unique destination (source) IP addresses inside the network in last N connections § Number of connections from source (destination) IP to the same destination (source) port in last N connections

Outlier Detection on DARPA’ 98 Data ROC curves for bursty attacks LOF approach is consistently better than other approaches ROC curves for single-connection attacks Unsupervised SVMs are good but only for high false alarm (FA) rate NN approach is comparable to LOF for low FA rates, but detection rate decrease for high FA LOF approach is superior to other outlier detection schemes Mahalanobis-distance approach – poor due to multimodal normal behavior Majority of single connection attacks are probably located close to the dense regions of the normal data

M I N D S Anomaly Detection on Real Network Data During the past few months various intrusive/suspicious activities were detected at the AHPCRC and at the U of Minnesota using MINDS Many of these could not be detected using state-of-the-art tool like SNORT Anomalies/attacks picked by MINDS w Scanning activities w Non-standard behavior § Policy violations § Worms

M I N D S Detection of Scans on Real Network Data August 13, 2002 w. Detected scanning for Microsoft DS service on port 445/TCP (Ranked #1) §Reported by CERT as recent Do. S attacks that needs further analysis (CERT August 9, 2002) §Undetected by SNORT since the scanning was non-sequential (very slow) Number of scanning activities on Microsoft DS service on port 445/TCP reported in the World (Source www. incidents. org) August 13, 2002 w. Detected scanning for Oracle server (Ranked #2) w Reported by CERT, June 13, 2002 w First detection of this attack type by our University w Undetected by SNORT because the scanning was hidden within another Web scanning

M I N D S Detection of Scans on Real Network Data w October 10, 200 §Detected a distributed windows networking scan from multiple source locations (Ranked #1) §Similar distributed scan from 100 machines scattered around the World happened at University of Auckland, New Zealand, on August 8, 2002 and it was reported by CERT, Insecure. org and other security organizations Attack sources Destination IPs Distributed scanning activity

Detection of Policy Violations on Real Network Data August 8, 2002 w Identified machine that was running Microsoft PPTP VPN server on non-standard ports, which is a policy violation (Ranked #1) w Undetected by SNORT since the collected GRE traffic was part of the normal traffic w Example of an insider attack October 30, 2002 w Identified compromised machines that were running FTP servers on non-standard ports, which is a policy violation (Ranked #1) §Anomaly detection identified this due to huge file transfer on a nonstandard port §Undetectable by SNORT due to the fact there are no signatures for these activities §Example of anomalous behavior following a successful Trojan horse attack

M I N D S Detection of Worms on Real Network Data w October 10, 2002 § Detected several instances of slapper worm that were not identified by SNORT since they were variations of existing warm code § Detected by MINDS anomaly detection algorithm since source and destination ports are the same but non-standard, and slow scan-like behavior for the source port § Potentially detectable by SNORT using more general rules, but the false alarm rate will be too high § Virus detection through anomalous behavior of infected machine Number of slapper worms on port 2002 reported in the World (Source www. incidents. org)

M I N D S MINDS - Framework for Mining Associations Anomaly Detection System Ranked connections attack Discriminating Association Pattern Generator normal update Knowledge Base 1. Build normal profile 2. Study changes in normal behavior 3. Create attack summary 4. Detect misuse behavior 5. Understand nature of the attack R 1: TCP, Dst. Port=1863 Attack … … R 100: TCP, Dst. Port=80 Normal

Discovered Real-life Association Patterns Rule 1: Src. IP=IP 1, Dst. Port=80, Protocol=TCP, Flag=SYN, No. Packets: 3, No. Bytes: 120… 180 (c 1=256, c 2 = 1) Rule 2: Src. IP=IP 1, Dst. IP=IP 2, Dst. Port=80, Protocol=TCP, Flag=SYN, No. Packets: 3, No. Bytes: 120… 180 (c 1=177, c 2 = 0) At first glance, Rule 1 appears to describe a Web scan Rule 2 indicates an attack on a specific machine Both rules together indicate that a scan is performed first, followed by an attack on a specific machine identified as vulnerable by the attacker

Discovered Real-life Association Patterns…(ctd) Dst. IP=IP 3, Dst. Port=8888, Protocol=TCP (c 1=369, c 2=0) Dst. IP=IP 3, Dst. Port=8888, Protocol=TCP, Flag=SYN (c 1=291, c 2=0) l l l This pattern indicates an anomalously high number of TCP connections on port 8888 involving machine with IP address IP 3 Follow-up analysis of connections covered by the pattern indicates that this could be a machine running a variation of the Kazaa file-sharing protocol Having an unauthorized application increases the vulnerability of the system

Discovered Real-life Association Patterns…(ctd) Src. IP=IP 4, Dst. Port=27374, Protocol=TCP, Flag=SYN, No. Packets=4, No. Bytes=189… 200 (c 1=582, c 2=2) Src. IP=IP 4, Dst. Port=12345, No. Packets=4, No. Bytes=189… 200 (c 1=580, c 2=3) Src. IP=IP 5, Dst. Port=27374, Protocol=TCP, Flag=SYN, No. Packets=3, No. Bytes=144 (c 1=694, c 2=3) …… l l This pattern indicates a large number of scans on ports 27374 (which is a signature for the Sub. Seven worm) and 12345 (which is a signature for Net. Bus worm) Further analysis showed that no fewer than five machines scanning for one or both of these ports in any time window

Discovered Real-life Association Patterns…(ctd) Dst. Port=6667, Protocol=TCP (c 1=254, c 2=1) l l This pattern indicates an unusually large number of connections on port 6667 detected by the anomaly detector Port 6667 is where IRC (Internet Relay Chat) is typically run Further analysis reveals that there are many small packets from/to various IRC servers around the world Although IRC traffic is not unusual, the fact that it is flagged as anomalous is interesting l This might indicate that the IRC server has been taken down (by a DOS attack for example) or it is a rogue IRC server (it could be involved in some hacking activity)

Discovered Real-life Association Patterns…(ctd) Dst. Port=1863, Protocol=TCP, Flag=0, No. Packets=1, No. Bytes<139 (c 1=498, c 2=6) Dst. Port=1863, Protocol=TCP, Flag=0 (c 1=587, c 2=6) Dst. Port=1863, Protocol=TCP (c 1=606, c 2=8) l l l This pattern indicates a large number of anomalous TCP connections on port 1863 Further analysis reveals that the remote IP block is owned by Hotmail Flag=0 is unusual for TCP traffic

M I N D S Conclusion Data mining based algorithms are capable of detecting intrusions that cannot be detected by state-of-the-art signature based methods w SNORT has static knowledge manually updated by human analysts w MINDS anomaly detection algorithms are adaptive in nature w MINDS anomaly detection algorithms can also be effective in detecting anomalous behavior originating from a compromised or infected machine MINDS Research u u u M I N D S Defining normal behavior Feature extraction Similarity functions Outlier detection Result summarization Detection of attacks originating from multiple sites Outsider attack u Network intrusion Insider attack u Policy violation Worm/virus detection after infection

Future Work Distributed Attacks coordinated from multiple locations Content Analysis Wireless Networks w No fixed infrastructure w Physical layer is less secure w No single check point M I N D S MINDS Collaboration M I N D S

Challenges of Wireless Networks Physical layer is less secure than in fixed computer networks Mobile nodes do not have fixed infrastructure There are no traffic concentration points where packets can be monitored There is no firewall no clearly defined protected perimeter There may be no clear separation between normal and anomaly, due to volatile physical movements

Intrusion Detection in Wireless Networks w. Threats in wireless networks w Eavesdropping – intruder is listening the data w Intrusions – intruder attempts to access and modify the data w Communication hijacking - a rogue node can capture the channel, may pose as a base station and seduce mobiles to connect to it and collect data (e. g. passwords, keys) and information from nodes w Jamming - disturbing the communication channel with various frequency domains and disabling all communication on the channel w. Wireless IDS cannot use the same architecture as network IDS w Multi-level IDS (incorporated in multiple layers of wireless networks) MINDS w Should run on each Collaboration mobile node w IDSs must cooperate w Should rely on anomaly detection M I N D S M I N D S

Wireless Networks in Army w U. S. Army recently announced the adoption of two wireless network systems for soldiers called "Land Warrior" and CAISI (Combat Automated Information System Interface) that provide wireless communication between the soldier and his leaders and support teams w Both wireless systems originally developed to be used with WEP(Wired Equivalency Privacy) and DES (Data Encryption Standard) w In 2001, it was demonstrated that WEP was flawed and insecure w In 1997, it was shown that DES is not secure w AES (Advanced Encryption Standard) based on Rijndael encryption algorithm that uses different key sizes w Air. Fortress. TM is a combination of hardware and software that attempts to provide security in wireless networks through sophisticated encryption, strong authentication and stringent access control w Still in development phase there is a need for wireless IDS

Data Mining in Commercial Word Given its success in commercial applications, data mining holds great promise for analyzing large data sets. Employed No NO Yes # of years <2 4 NO # of years in school 2 Yes >4 YES Classification / Predictive Modeling {Direct Marketing, Fraud Detection, Credit Risk Analysis} Clustering (Market segmentation) Association Patterns Marketing / Sales Promotions Back