Mobile crowd computing with mobile agents and work stealing

Abstract

Mobile Crowd Computing (MCC) is a computing paradigm that leverages the collective power of idle smart mobile devices to perform computationally demanding tasks. Since a mobile crowd consists of diverse smart mobile devices (i.e. varying processing capacities, memory, battery, etc.) load balancing becomes an important aspect to achieve expected outcomes. We hypothesize that effective load balancing for MCC can be achieved using a combination of mobile agents and a work stealing approach. Thus, in this paper, we present the design, implementation, and evaluation of our mobile agent based MCC framework which uses work stealing for load balancing.