We are working on methodologies and techniques to manage huge amounts of monitoring data (OSNs, p2p, high speed links, ...) from both active and passive approaches.
Network data compression
We are developing a compressed representation technique for network monitoring data that provides:
- reduced memory footprint for stored data, trading off accuracy versus space occupation;
- a wide class of operations allowed directly on compressed data with no need for decompression and with reduced time complexity.
The research has been partially conducted during a summer internship by Mr. Giuseppe Aceto at DOCOMO Innovations Inc. (formerly Docomo Communications Laboratories USA), Palo Alto, CA.
Network data reduction
We are working on an entropy-based methodology to reduce network traffic data which provides several advantages:
- there is no need to consider the entire data sets in the processing stage;
- allows to work off-line on reduced data.
