A Distributed Allocation Strategy for Data Mining Tasks in Mobile Environments

The increasing computing power of mobile devices has opened the way to perform analysis and mining of data in many real-life mobile scenarios, such as body-health monitoring, vehicle control, and wireless security systems. A key aspect to enable data analysis and mining over mobile devices is ensuring energy efficiency, as mobile devices are battery-power operated. We worked in this direction by defining a distributed architecture in which mobile devices cooperate in a peer-to-peer style to perform a data mining process, tackling the problem of energy capacity shortage by distributing the energy consumption among the available devices. Within this framework, we propose an energy-aware (EA) scheduling strategy that assigns data mining tasks over a network of mobile devices optimizing the energy usage. The main design principle of the EA strategy is finding a task allocation that prolongs network lifetime by balancing the energy load among the devices. The EA strategy has been evaluated through discrete-event simulation. The experimental results show that significant energy savings can be achieved by using the EA scheduler in a mobile data mining scenario, compared to classical time-based schedulers. Carmela Comito DEIS, University of Calabria, Rende (CS), Italy, e-mail: [email protected] Deborah Falcone DEIS, University of Calabria, Rende (CS), Italy, e-mail: [email protected] Domenico Talia ICAR-CNR and DEIS, University of Calabria, Rende (CS), Italy, e-mail: [email protected] Paolo Trunfio DEIS, University of Calabria, Rende (CS), Italy, e-mail: [email protected]