Efficient and scalable execution of smart city parallel applications

摘要

Recent research efforts in the field of urban computing aim to develop innovative services for citizensrnthrough the application of ubiquitous and pervasive computing paradigms in urban spaces.rnSmart city applications need to cope with a large number of involved users and devices. Since datarnand objects are strictly related to the territory on which they are defined and used, it is preferable,rnwhen possible, to perform computation locally through the adoption of dispersed computingrnnodes such as CPU-equipped sensors. In this context, the computation related to smart city applicationsrncan be profitably and efficiently parallelized by partitioning the territory into regions andrnassigning the computation related to each single region toa local node.Nevertheless, the adoptionrnof parallel computingmodels poses several communication and synchronization issues, especiallyrnwhen the number of nodes is large and the time constraints of applications are compelling. Thisrnpaper presents and analyzes a parallel computingmodel for smart city applications in which eachrnnode needs to exchange information only with a subset of neighbor nodes, allowing the synchronizationrnoverhead to be significantly reduced. As sample application, we consider the analysisrnand prediction of internet traffic generated by vehicle and pedestrian devices moving on a smartrnavenueequippedwith distributed computing nodes. Thisworkoffers adetailedperformance evaluationrnin a number of scenarios, including uniform and nonuniform user distribution and differentrntypes of user mobility behavior. The results show that the presented computation model offersrnnotable advantages in terms of computation efficiency and speedup, with respect to a classicalrnall–to-all synchronization paradigm, in which the nodes need to coordinate with a central entity.