Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Network traffic based computer system user identification Dr Zsolt Illési associate professor College of Dunaújváros [email protected] Open Source Intelligence Areas of Development Key Questions of Detection Who? (individual(s) involved) How? (used tools/ exploits) When? (timeline) Provide Information About What? (nature of events) Why? (motivation) Where? (scene) Open Source Intelligence Areas of Development Time to repair (TTR) or 'downtime' Incident Lifecycle Response time Restoration time Detection IDh IDu Diagnosis Recovery Restoration IDo Time Incident Incident Recovery time Uptime (TBL-TTR) Time between systems incidents (TBF) Correction Initiation Diagnosis Open Source Intelligence Areas of Development Remedy Network Situational Awareness • Cyber Attack Scenarios • Situation-Aware and Context-Aware Network Applications • CERTs and CSIRTs • Security Event and Information Management • Application Security, Audits and Penetration Testing Open Source Intelligence Areas of Development Web Traffic Characterisation • • • • Intrusion Detection Systems Traffic Characterisation Techniques Web Analytics Security Incident Response Open Source Intelligence Areas of Development Cyber Situational Awareness Tools & Techniques • Fuzzy Logic • Rough Set • Artificial Neural Networks • Artificial Intelligence • Genetic Algorithm • Evidence Theory (DST) • Bayesian Networks & Set Theory • Big Data Analytics • Game Theory • Graph Theory Open Source Intelligence Areas of Development Identifying someone • Prove that a signature is from a known person • Prove that some network traffic is generated by a specific user Open Source Intelligence Areas of Development Bayesian interpretation of network data posterior knowledge new data prior knowledge 𝑃(𝐻0 |𝐸) 𝑃(𝐸|𝐻0 ) 𝑃(𝐻0 ) = ∙ 𝑃(𝐻1 |𝐸) 𝑃(𝐸|𝐻1 ) 𝑃(𝐻1 ) posterior odds likehood ratio Open Source Intelligence Areas of Development prior odds Identification of WHO using a computer? (Assumptions) • User(s) in action – one or more person – one or more computer system – carefully defined (limited) task performance • Used network data – generic protocol data are available – payload (e.g. data) possibly encrypted • Previous information (reference data model is available) Open Source Intelligence Areas of Development Identification of WHO using a computer? (Tools) • Network taps (specialised hardware or active network tool) • Sniffers, and network traffic/data analysers (wireshark, tcpdump, tcpstat, tcptrace, CoralReef etc.) • Scripting language for data pre-processing (Python, Pearl etc.) • Number cruncher (Octave, Scilab, Matlab, Mathematica etc.) Open Source Intelligence Areas of Development Identification of WHO using a computer? (Stages) • Reference data network usage data collection (prior probability distribution) • Definine the probability that a certain person (or computer system) uses the network (hypothesis testing; posterior distribution analysis) Open Source Intelligence Areas of Development Identification of WHO using a computer? (Process) • Raw network data collection • Understand network data ( – packet sorting and analysis – data-flow and protocol statistics – network connection (source-destination pairing) • Bayesian analysis (current data vs reference data) Open Source Intelligence Areas of Development Pro’s and Con’s Constraints Benefits • Single user • 80%+ accuracy (pls consider the • No other (significant) limitations!!!) interference to computer traffic (e.g background software activity) • lack of adequate amount of reference data (directed network usage) Open Source Intelligence Areas of Development Future development — Experiment Scope • Greater reference data – number of persons – duration of network usage – mixed data with some other subjects • Combine with logs (apply the results to log analisis fileld and enhance accuracy) Open Source Intelligence Areas of Development Future Developments — Combined approach • • • • • • • • Hidden Markov model Gaussian mixture models Fuzzy Logic Artificial Neural Networks Data Mining Decision Trees Graph Theory etc. Open Source Intelligence Areas of Development