Presentation is loading. Please wait.

Presentation is loading. Please wait.

Mitigating the Insider Threat using High-dimensional Search and Modeling Presenter: Eric van den Berg Wednesday, July 13, 2005.

Similar presentations


Presentation on theme: "Mitigating the Insider Threat using High-dimensional Search and Modeling Presenter: Eric van den Berg Wednesday, July 13, 2005."— Presentation transcript:

1 Mitigating the Insider Threat using High-dimensional Search and Modeling Presenter: Eric van den Berg evdb@research.telcordia.com Wednesday, July 13, 2005 Team: Shambhu Udpadhyaya, Hung Ngo (SUNY Buffalo) Muthu Muthukrishnan, Raj Rajagopalan (Rutgers) DARPA IPTO Program: Self Regenerative Systems (SRS) Program Manager: Lee Badger PI: Eric van den Berg

2 SRS PI meeting July 2005 – 2 Project overview  Project goal: to build a system that defends critical services and resources against insiders, which – Correlates large numbers of sensor measurements – Synthesizes appropriate pro-active responses  What is done today? – Reactive systems: Detect attacks late in cycle – Anomaly detection systems: Few streams for correlation – Human-based systems: not scalable – Collateral damage may be large

3 SRS PI meeting July 2005 – 3 Project overview (continued)  Technical Approach – Large network of sensors, to let insider trigger alerts – High dimensional network state description using sensor alerts – Search engine finds top-K past states similar to sensor snapshot – Insider modeler and analyzer tool used to identify attack points, train search engine, guide sensor placement – Response engine to analyze impact on critical services and synthesize reconfiguration response  Technical Challenges – Testing SVD-based search technology in a new domain – New ‘Insider analyzer’ key-challenge graph problem is hard – Training search engine, labeling and annotating states

4 SRS PI meeting July 2005 – 4 Project overview (continued)  Quantitative Metrics to measure success and overheads – False alarm / detection rate – Test detection for novel variations of known attacks  Major Achievements to date – Initial prototype for sensor network – Initial prototype for SVD-based search engine – Initial prototype for Insider modeler and analyzer tool – First test with independent ‘insiders’ Task(milestone) Jan-Jun 05 Jul-Dec 04 Design (document) Prototyping (software) Testing (report) Jul-Dec 05

5 SRS PI meeting July 2005 – 5 Architecture

6 SRS PI meeting July 2005 – 6 Insider analyzer and modeler  Insider threat manifests in two forms: – Insider abuse while staying within legitimate privileges – Insider abuse while exceeding assigned privileges  Focus on an insider's view of an organization: hosts, reachability and access control  A new threat model called a “key challenge graph” – Similar to attack graphs, less emphasis on details – Allows static analysis of insider threat  More in papers Threat analysis metricCost of Attack Actual targetTarget Vertex Location of insiderStarting Vertex Access ControlKey Challenge Information, CapabilityKey Connectivity, ReachabilityEdge Hosts, PeopleVertex AbstractionModel Component

7 SRS PI meeting July 2005 – 7 Insider modeler and analyzer MAPIT tool architecture Network entity rules Cost Rules MAPIT Engine Network topology Key challenge graph Vulnerabilities Authentication mechanism Social Eng. Awareness Sensitivity analysis Defense centric analysis

8 SRS PI meeting July 2005 – 8 Example physical network: Hackfest 2004

9 SRS PI meeting July 2005 – 9 Key challenge graph: logical network

10 SRS PI meeting July 2005 – 10 Sensors to detect insider attacks  Detect changes from user ‘normal behavior’ – Profile anomaly detector – Statistical sequential change point detection – Future: biometrics, e.g. keystroke dynamics?  Detect access to target resources – Pluggable Authentication Module, File integrity checker  Other useful sources: – web, audit logs (e.g. internal website searches) – network intrusion detectors (signature, anomaly)

11 SRS PI meeting July 2005 – 11 Network traffic anomaly detector  Streaming data model – Large data volume and speed: in backbone 1 billion packets/hour/router – Large data domain: IPv4: 2^32 addresses, IPv6: 2^128  Consequences: – Can scan data (at most) once – Need small-space structure to summarize data  Hard to store O(n) data points when n=2^32  Cannot store at 2^128  Idea: build synopsis data structure for IP-packets – CM-sketches, deltoid group-testing  Detect attacks based on changes in traffic volume – Currently: traffic to destination IP address (likely targets)  Can detect attacks exhibiting large changes in packet distribution

12 SRS PI meeting July 2005 – 12 Example: Network anomaly detector  Based on week 2 of 1999 MITLL data – from inside sniffer  Traffic volume based anomaly detection – Ipsweep, portsweep, phf, httptunnel, etc.  Detects targets of all four above attacks – Does give additional big changes ~1%, not attacks  Search engine to filter out non-attacks

13 SRS PI meeting July 2005 – 13 Sensor alert message format  We use IDMEF (Intrusion Detection Message Exchange Format) to transmit and store sensor alerts – Between sensors and database – Between search engine and response engine  Alert storage in mySQL database with IDMEF- based schema

14 SRS PI meeting July 2005 – 14 Network state description  Network state is constructed from sensor alerts: – Accommodate heterogeneous sensor types – Account for different sensitivity of sensor types – Tolerate possibly delayed or missing, ‘out of order’ alerts  Alerts are mapped to a high-dimensional vector for search – Coordinates correspond to different sensor-alert types – Some possibilities for mapping values:  Total number of sensor alerts of given type in (sliding) time window  Indicator: sensor alert occurred in (sliding) time window  Network state is labeled: – With Classification e.g. ‘Normal’, ‘Insider’ – With Response for Response Engine

15 SRS PI meeting July 2005 – 15 High-dimensional search engine  Goal: Find historical documented network states most similar to the current network state snapshot  Output: Top-K list of ranked/prioritized similar states  Ranking can be based on similarity metric, or – potential impact, e.g. attack ‘risk’  Impact of historical network states is documented, – impact of current state analyzed with Response engine  Search engine reduces search space dimensionality – Using Singular Value Decomposition, or random projection  Similar states found by nearest neighbor search – distance metric: e.g. cosine similarity, Euclidean distance

16 SRS PI meeting July 2005 – 16 Ranking via alert correlation  Combine alert information from network and host sensors  Segment alert state vector to reflect activity by host and user  Reinforce or weaken ‘attack’ hypothesis  Useful as component to detect or visualize specific attack patterns (moving from host to host)

17 SRS PI meeting July 2005 – 17 SVD-based anomaly detection  Statistical Methods using ideas from Principal Component Analysis (PCA) – Imagine alarm vectors come from multivariate normal distribution – Compute sample mean, covariance / correlation matrix for training data – Eigenvalue decomposition of covariance matrix to separate data into normalized independent components

18 SRS PI meeting July 2005 – 18 anomaly detection (cont.)  Test new vector of alarms – Check for alarms not in training data – Check for fit to training distribution  Status – Code ready – Still to determine thresholds  How far to use normality assumptions vs. switching to nonparametric methods

19 SRS PI meeting July 2005 – 19 Early detection of insider attacks  How to represent time evolution in multi-stage attacks?  Like learning attacks from documented historical network states, we can also document attack precursors or attack stages – Full attack now represented as a sequence of network state vectors – Robust against slow attacks: no explicit dependence on time – Would like to make ‘precursor’ annotation (semi-) automatic  Approaches to automatic precursor annotation – Temporal precursors – Spatial precursors

20 SRS PI meeting July 2005 – 20 Schematic for using precursors

21 SRS PI meeting July 2005 – 21 Impact Analysis using Response Engine  Building upon Smart Firewalls technology from Dynamic Coalitions program; Response Engine – Has overview of current network configuration – Logically validates Policies, expressed in terms of end-to-end service availability – Generates candidate reconfigurations to comply with Policies as much as possible  In this project – Detected attack type and location is translated into its effect on the stated policies and current network configuration – E.g. Server failure due to a Denial of Service attack  Response Engine can analyze the impact of both the attack and its candidate responses on the availability of critical resources – E.g. Analyze impact of vulnerability exploit: how widespread is the vulnerability?  Administrator can push response into the network

22 SRS PI meeting July 2005 – 22 Response using policy-based architecture Policy Response Engine Topology High-level Policy Configuration Summarized Configuration Routers & Switches Control & Monitor Device-level Policy Configuration Detailed Configuration Security Policy Adaptors Correlated Alerts

23 SRS PI meeting July 2005 – 23 First test results  First system test by independent insiders  Goal: extract operations-sensitive military information  Four volunteer ‘insiders’ given – existing account information – starting location and – nature of target  Result: 3 out of 4 attackers detected  Program goal: delay / thwart 10% of insider attacks

24 SRS PI meeting July 2005 – 24 Next / future steps  Show effectiveness against wide range of attacks  Measure false positive rate  Adapt detection system to heterogeneous environments

25 SRS PI meeting July 2005 – 25 Backup slides

26 SRS PI meeting July 2005 – 26 SVD-based search on Test alert set  Attacks from MITLL Scenario Specific datasets  Alerts from NC-State (TIAA) – Generated by Real-Secure IDS on MITLL attack data sets  Scenario 1: 1. IP sweep 2. Probe for sadmind vulnerability 3. Break-in via sadmind exploit 4. Installation of mstream DDoS software 5. Launch DDoS attack  Scenario 2: 1. Probe of DNS server via HINFO query 2. Break-in via sadmind exploit 3. FTP upload of mstreamDDoS software and attack script 4. Initiate attack on other hosts 5. Launch DDoS attack

27 SRS PI meeting July 2005 – 27 Top-7 most similar states, lower = more similar

28 SRS PI meeting July 2005 – 28 Most similar states and attack phases

29 SRS PI meeting July 2005 – 29 Scenario 2 tested on Scenario 1 history

30 SRS PI meeting July 2005 – 30 Most similar states in Scenario 1


Download ppt "Mitigating the Insider Threat using High-dimensional Search and Modeling Presenter: Eric van den Berg Wednesday, July 13, 2005."

Similar presentations


Ads by Google