1 / 49

Intrusion Detection Alert Correlation

Intrusion Detection Alert Correlation. Mark Shaneck 2/11/2005. Outline. Problem Statement Different Correlation Approaches A Comprehensive Approach Good News and Bad News A Better Approach?. What’s The Problem?. Large organizations get tons of alerts Possibly up to 20,000 per day!

Télécharger la présentation

Intrusion Detection Alert Correlation

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Intrusion Detection Alert Correlation Mark Shaneck 2/11/2005

  2. Outline • Problem Statement • Different Correlation Approaches • A Comprehensive Approach • Good News and Bad News • A Better Approach?

  3. What’s The Problem? • Large organizations get tons of alerts • Possibly up to 20,000 per day! • Many false alarms

  4. Also… • Alerts can come from many different sources • Signature based IDS (Snort) • File System Integrity Checkers • System Call Traces • Alerts may represent multiple stages in one attack • Hard to make sense out of a large pile of alerts!

  5. So What Is Alert Correlation? • 3 general categories • Alert Clustering • Matching Predefined Attack Scenarios • Prerequisites/Consequences

  6. Alert Clustering • Main Sources: • A. Valdes, K. Skinner, “Probabilistic Alert Correlation”, RAID 2001 • O. Dain, R. Cunningham, “Building Scenarios from a Heterogeneous Alert Stream”, IEEE Workshop on Information Assurance and Security, 2001

  7. General Idea • Join alerts together in some meaningful groups • Group alerts into attack threads - one thread contains all alerts related to one attack • For a new alert, compare to all alert threads • Join to the closest match • Or start new thread if none match

  8. Similarity Measure • Feature Overlap - only consider features present in both (source, target, ports, attack class, timestamps, etc.) • Each feature has a similarity measure • How much do port lists overlap? • Is one port contained within another’s list? (target port was previously scanned) • Are the IPs from the same subnet? • Attack classes have a similarity matrix

  9. Similarity Expectation • Different levels of similarity are expected for different features in different situations • SYN FLOOD with source spoofed • Expectation of similarity for source IP is 0 • Scanning port(s) • Expectation of target IP is low (but not 0 - since it usually scans the subnet)

  10. Minimum Similarity • Threshold for similarity measure • Similarity is 0 if not above the minimum • Adjusting thresholds • Synthetic Threads • high for sensor id, IPs • Security Incidents • low for sensor id, high for attack class • fuse alerts from multiple sources • Multistep attack detection • low for attack class

  11. So What Is Alert Correlation? • 3 general categories • Alert Clustering • Matching Predefined Attack Scenarios • Prerequisites/Consequences

  12. Matching Predefined Attack Scenarios • Main sources • H. Debar, A. Wespi, “Aggregation and Correlation of Intrusion-Detection Alerts”, RAID 2001 • B. Morin, H. Debar, “Correlation of Intrusion Symptoms : an Application of Chronicles”, RAID 2003

  13. Aggregation and Correlation • Correlation • Group alerts that are part of the same attack trend • Duplicates • Consequences (chain of related alerts) • Aggregation • Group alerts based on certain criteria to aggregate severity level, reveal trends, clarify attacker’s intentions • Situations

  14. Duplicates • Duplicates Definition • Initial Alert Class • Duplicate Alert Class • List of Attributes (that must be equal) • Severity Level (new severity level for new merged alert) • Specified by analyst

  15. Consequences • Consequences Definition • Initial Alert Class • Initial Probe Token • Consequence Alert Class • Consequence Probe Token • Severity Level • Wait Period • Links together alerts that are sequential in nature

  16. Aggregation • Aggregate based on three axes • Alert Class • Source • Target • Putting wildcards for different cases gives different views • Aggregate into scenarios

  17. Scenarios • Same source/target/attack class • A single attacker launching attacks against a single victim • Same source/destination • Single attacker running many attacks on a single victim • Same target/attack class • Distributed attack against a single victim • Same source/attack class • A single attacker running the same attack against multiple victims

  18. Chronicles • “Set of events, linked together by time constraints, whose occurrence may depend on the context” • Similar to plan recognition • Used to model known attack “chunks” • Long attack scenarios may have many paths • Certain small sequences of events almost certainly occur together

  19. So What Is Alert Correlation? • 3 general categories • Alert Clustering • Matching Predefined Attack Scenarios • Prerequisites/Consequences

  20. Prerequisites/Consequences • F. Cuppens, A. Miège, “Alert Correlation in a Cooperative Intrusion Detection Framework”, In IEEE Symposium on Security and Privacy, 2002 • P. Ning, D. Reeves, et al. (many papers) • Check my website for the list • Or the very last slide…..

  21. Prerequisites/Consequences • Prerequisite: the necessary condition for the attack to be successful • Consequence: the possible outcome of the attack • Represented as a logical formula • Using only AND and OR connectives

  22. Hyper Alert Type • (fact, prerequisite, consequence) • SadmindBufferOverflow = ({VictimIP, VictimPort}, ExistHost(VictimIP) AND VulnerableSadmind(VictimIP) {GainRootAccess(VictimIP)})

  23. Prepare-For Relationships • An alert “prepares for” another alert if it contributes to the second alert’s prerequisite set • Also must occur earlier in time

  24. Correlation Graph • Directed acyclic graph, with the nodes being alerts and the edges being the prepares-for relations • Could be huge!

  25. Adjustable Reduction • Aggregation of alerts of the same type • Can result in overly simple graphs • Adjustable • Analyst can specify a time interval • Only alerts with time gap less than the interval are merged

  26. Adjustable Reduction

  27. Focused Analysis • Logical combination of comparisons between attribute names and constants • SrcIP = 129.174.142.2 OR DestIP = 129.174.142.2 • Useful for focusing on a critical server

  28. Graph Decomposition • Cluster alerts based on “common” features • Use clusters to separate large graph into smaller ones • (A1.SrcIP = A2.SrcIP) AND (A1.DestIP = A2.DestIP) • Clustering constraints are specified by the analyst

  29. Reduced and DecomposedGraph Example

  30. Matching Attack Strategies • Attack Strategy Graph • Set of events linked together by certain constraints • Time Order • IP Addresses • Events can be generalized to deal with variations SadmindBufferOverflow RPCBufferOverflow TooltalkBufferOverflow

  31. Measuring Similarity Between Attack Strategies • Error Tolerant Graph Isomorphism • Use edit distance to derive a similarity measure • Can be used to find similar attacks or to match against predefined strategies

  32. Hypothesizing About Missed Attacks • Missed attacks can break up the graphs • One attack graph becomes two disconnected, seemingly unrelated, attack graphs • Indirect Prepares-for • Similarity based merging of attack graphs • Prune hypotheses with network traffic • E.g. one hypothesized attack is ICMP ping, but no ICMP traffic occurred during that time

  33. Outline • Problem Statement • Different Correlation Approaches • A Comprehensive Approach • Good News and Bad News

  34. A Comprehensive Approach • F. Valeur, G. Vigna, C. Kruegel, R. Kemmerer, "A Comprehensive Approach to Intrusion Detection Alert Correlation", In IEEE Transactions on Dependable and Secure Computing, 2004

  35. Alert Fusion • Combine alerts that are independent detection of the same attack instance • Must be temporally close • From different sensors • Identical overlapping attributes

  36. Alert Verification • Idea: False positives can negatively impact alert correlation • Filter out false positives and irrelevant positives (alerts that correspond to failed attacks)

  37. Alert Verification • Passive: use network knowledge to see if attack could succeed (low overhead, low confidence) • Listing of existence of/services running on IPs • Firewall configurations • Active: check for evidence (high overhead, high confidence) • See if service is still running and available • See if extra ports are open • Use vulnerability scanner to test target machine • Remote login and run scripts

  38. Thread Reconstruction • Group alerts that refer to attacks launched by one attacker against a single target • Merge alerts with same source and destination and within a time interval

  39. Attack Session Reconstruction • Link network based alerts to host based alerts • Manually specify links between network events and process events • Alert on web server process (or one of its children) can be correlated to a (temporally) nearby network alert targeted to that machine on port 80

  40. Focus Recognition • Identify hosts that are the source or target of lots of attacks • Merge these alerts together into one • Source: Scanning • Target: DDoS

  41. Multi-Step Correlation • Identify attack patterns that are made up of multiple individual attacks • Create attack patterns by means of expert knowledge • Simply match the merged alerts to the attack strategies

  42. Experimental Results • Defcon9 • Input: 6,378,096 alerts • Output: 203,303 alerts • Reduction: 96.81% • TreasureHunt • Input: 2,811,169 alerts • Output: 1,080 alerts • Reduction: 99.96% • MIT/LL 2000 • Input: 36,635 alerts • Output: 17,220 • Reduction: 53.00%

  43. Benefits of Alert Correlation • Higher level representation of alerts reduces clutter and can show attack structure • Reduce false positives • False positives are unlikely to correlate with other alerts • May find many attacks and respective scenarios

  44. Limitations of Correlation • Relies on IDS to alarm each step of the attack • Exploit mutations • Novel attacks • Bad sensor placement • Sensor overload - packet loss • Restricted ruleset for better performance • Relies heavily on a priori expert knowledge

  45. Limitations of Correlation (cont) • Cannot provide a comprehensive view on network attacks

  46. MINDS Level 2 • Level 1 IDS alerts • Anchor Point Identification • Context Extraction • Attack Characterization • Behavior/Host Profiling

  47. Questions? • Paper links located at: http://www.cs.umn.edu/~shaneck/wormlist.html • At the bottom of the page • Slides available: http://www.cs.umn.edu/~shaneck/Correlation.ppt

  48. A Budding Hacker

  49. Peng Ning Reference List • P. Ning, D. Reeves, Y. Cui, "Correlating Alerts Using Prerequisites of Intrusions", Technical Report, TR-2001-13, North Carolina State University, Department of Computer Science, December 2001 • P. Ning, Y. Cui, D. Reeves, "Analyzing Intensive Intrusion Alerts via Correlation", In Recent Advances in Intrusion Detection, 2002 • P. Ning, Y. Cui, D. Reeves, "Constructing Attack Scenarios through Correlation of Intrusion Alerts", In CCS 2002 • P. Ning, D. Xu, "Learning Attack Strategies from Intrusion Alerts", In CCS 2003 • P. Ning, D. Xu, C. Healey, R. St. Amant, "Building Attack Scenarios through Integration of Complementary Alert Correlation Methods", NDSS, February 2004 • Y. Zhai, P. Ning, P. Iyer, D. Reeves, "Reasoning about Complementary Intrusion Evidence", 20th Annual Computer Security Applications Conference, December 2004 • D. Xu, P. Ning, "Alert Correlation Through Triggering Events and Common Resources", 20th Annual Computer Security Applications Conference, December 2004 • P. Ning, D. Xu, "Hypothesizing and Reasoning about Attacks Missed by Intrusion Detection Systems", ACM Transactions on Information and System Security, 2004

More Related