Mobile network traffic data compression by means of wavelet decomposition

Third generation mobile networks rely on an architecture where a certain number of operation and maintenance centers handle a large amount of information about the network behavior. In this scenario, the amount of traffic-related data may become so large that even straightforward processing tasks, such as off-line querying, become a serious challenge. In this paper we show how, by integrating different techniques for database performance improvement, we may efficiently compress databases containing hundreds of gigabytes of mobile network traffic data. Our application relies on a wavelet-based synopsis technique that strongly reduces database size and allows fast database access, while preserving query syntax and semantic, at the price of a bounded information loss. Our solution has been tested on data coming from the Chinese and Italian mobile networks.