‘Work stealing’ is a good candidate for this since it possesses the aforementioned characteristics, and the aim of this paper is to present a work stealing approach to dynamically load balanced mobile crowd computing. The inherent problems of mobile computing such as resource scarcity, finite energy and low connectivity pose problems for most applications. These problems can be addressed by ‘sharing’ resource intensive work with are source rich server. We define ‘mobile crowd computing’ as a local ‘mobile resource cloud’ comprising of a collection of local nearby mobile devices, utilized to achieve a common goal in a distributed manner. Work distribution in a mobile environment poses a different set of issues than a typical distributed/grid environment: 1) Less processing power on a mobile device than on s node in distributed processing system. 2) A mobile node has on a finite energy source. 3) A resource pool made up of mobile devices is highly volatile, and hence node availability is inconsistent. 4) In a mobile cloud, the devices will be unknown to each other a priori, unlike in a grid environment where nodes are established and approved beforehand. Therefore a mobile cloud calls for a more opportunistic and ad hoc behavior. 5) A mobile cloud is most likely to be heterogeneous