A Multi-broker Cloud Architecture for the Purpose of Large Scale Sensing Applications Development

In this paper, we propose a multi-broker cloud architecture for the purpose of large scale sensing applications development. A set of heterogeneous Sensor cloudlets are deployed everywhere to collect data and disseminate them to a set of known sink nodes. An Unmanned Aerial Vehicle (UAV) cloud is used for the purpose of data muling from sink nodes to data centers. In this work, we address the problem of optimal deployment of UAVs over the sky to serve the maximal possible number of incoming requests, maximize the connectivity between busy UAVs, and minimize the consumed UAV energy. A set of algorithms are provided for the re-deployment of idle UAVs, and the reallocation of data collection tasks of UAVs. A simulation is conducted to evaluate the proposal.