Learn, Adapt, and Profile (LeAP):

Beating the odds in traffic measurements/detection with optimal online learning and adaptive policies

Traffic Profiles

A key tool for understanding and engineering Internet backbone is the analysis of packet traces. However, given the increasing backbone speed towards 100Gbps, it is prohibitive to monitor individual flows at all times. This project develops optimal online learning and adaptation strategies for accurate traffic sampling, inference, and detection under hard resource constraints (e.g., limited CPU or memory at routers) and dynamic network/traffic conditions. Based on theories and techniques in multi-arm bandits, group testing, and compressed sensing, optimal or near-optimal solutions will be developed by exploiting the unique structures of the specific measurement application under study. Challenges addressed include learning from observations with heavy-tailed distributions and long-range dependencies, coping with sparse and/or imperfect observations, and distributed learning strategies that involve multiple monitors and decision points.

If successful, this research will provide fundamental design principles for a flexible traffic measurement infrastructure under the software-defined networking (SDN) paradigm. Reconfigurable measurements based on a learning process can be realized in commodity router/switches using SDN APIs such as OpenFlow, leading to potential development of new services.




Graduate Students

  • Mehdi Malboubi, ECE (PhD)
  • Chao Wang, ECE (PhD)
  • Shu-Ming Peng, ECE (MS)
  • Lingxuan Li, ECE (MS)


  • Liyuan Wang, ECE (MS, Dec 2013)


M. Malboubi, L. Wang, C-N. Chuah, and P. Sharma, "Intelligent SDN based Traffic (de)Aggregation and Measurement Paradigm (iSTAMP)," IEEE INFOCOM, April/May 2014.

M. Malboubi, C. Vu, C-.N. Chuah, and P. Sharma, "Compressive Sensing Network Inference with Multiple-Description Fusion Estimation," IEEE Globecom, December 2013. [pdf]

L. Wang, Adaptive Network Traffic Estimation using OpenFlow: An Implementation in Mininet, MS Thesis, UC Davis, December 2013.


This project is supported by National Science Foundation CNS-1321115 grant (2013-2016) and HP Labs 2013 Innovation Research Award.