Scheduling of Distributed Applications in HHPCaaS Clouds for Internet of Things

The High Performance Computing as a Service provides the computing power to client applications on a pay per-use basis. Contemporary high-performance servers gain high computing power by applying multi-core processors, graphics processors and FPGA accelerators. But efficient utilization of computing resources requires adequate management methods. In this paper, we propose a new method of scheduling IoT distributed applications in the heterogeneous cloud environment. We assume that applications, which have been requested for execution, are specified as directed acyclic graphs. Our method dynamically schedules all tasks using resource sharing by the applications. The goal of scheduling is to minimize the cost of resource hiring and to minimize the latency of all activated applications. Experimental results showed that our method gives significantly better utilization of computational resources than existing management methods for clouds. Thus, the cost of using the cloud can be reduced.