Network Traffic Measurements and Analysis (Phd Course - 20/21)
This course will focus on fundamental aspects of large-scale network traffic measurements and analysis techniques, targeting two main challenges: (i) how to measure a network and (ii) how to extract knowledge from network traffic measurements. This course will start off by overviewing the most popular network measurement techniques at different layers of the communication stack: ranking from passive probes at the physical layer for wireless networks up to the Simple Network Management Protocol (SNMP) at the application layer.
Building upon these tools, the course will then focus on several use cases, overviewing for each one the main building blocks of the measurement system and the specific machine learning and data processing algorithm that can be used to extract knowledge from the acquired data.
Some of the use cases are the following:
- Anomaly detection and traffic identification
- Load and performance prediction in mobile radio networks
- User localization, behavior estimation, device profiling and classification in WiFi networks
Depending on the specific application scenario, supervised and unsupervised machine learning approaches will be studied and applied. The course will include hands-on lectures to implement network measurement and analysis systems.
Tue, Nov 17, 09:30 - 13:30 - Introduction, active measurements - slides
Wed, Nov 18, 09:30 - 13:30 - Passive measurements, data visualization techniques - visualization examples, RTT-distance script, server csv
Thu, Nov 19, 09:30 - 13:30 - Supervised machine learning - slides, code examples
Tue, Nov 24, 09:30 - 13:30 - Unsupervised learning - slides, code examples
Wed, Nov 25, 09:30 - 13:30 - Traffic classification and anomaly detection - slides 1, slides 2, code example
Thu, Nov 26, 09:30 - 13:30 - WiFi sniffing, project assignments - slides 1, slides 2