Mobile Interactions and Computation Offloading in Drop Computing

In recent years, the amount of data consumed by mobile devices has grown exponentially, especially with the advent of the Internet of Things and all its connected devices. For this reason, researchers are looking for methods of alleviating the congestion and strain on the network, generally through various means of offloading, or by bringing the data and computations closer to the devices themselves through edge and fog computing. Thus, in this paper we propose an extension to the Drop Computing paradigm, which introduces the concept of decentralized computing over multilayered networks. We present a novel offloading technique to be employed by Drop Computing nodes for increasing processing speed, reducing deployment costs and lowering mobile device battery consumption, by using the crowd of mobile nodes belonging to humans and the edge devices as opportunities for offloading data and computations. We compare our method with the initial Drop Computing implementation and with the default scenario for mobile applications and show that it is able to improve the overall network performance. We also perform an analysis of human interactions with two monitoring nodes located in an academic environment, to obtain realistic data and to extract behavior patterns regarding human habits and interactions, that aid us in developing an efficient offloading solution.