Optimizing Cost of Irregularly Structured Problems in Cloud

Abstract

This paper deals with maximizing the cost of parallel applications in a cloud-based environment. The cost belongs to monetary cost and cost of efficiency. The core argument seems to be that the parallel program robustness should affect the current monetary cost. Dynamic method of optimization is used to minimize the cost of computation. In order to measure the overall monetary cost of parallel computation, a cost model is used to evaluate the cost of parallel infrastructure as well as the cost of delayed performance. The main purpose of this cost model is to identify the necessary resources for performing this type of operation. Different methods have been used in the cloud environment. But these solutions do not take into account the uncertainties in the scheduling system, namely task start / perform / finish time, the unpredictable data transfer period between tasks, the unexpected arrival of new tasks. Such factors contribute to the breach of the task deadline and increase the cost of renting the service of executing the task, this effect will increase the monetary cost. Will boost the output by reducing the ambiguity in the scheduling process that requires time for execution of tasks and time for data transfer. In order to be precise a scheduling algorithm, uncertainty-Aware Scheduling Algorithm (ASA) is built to schedule complex and multiple tasks. When a task has been accomplished, its beginning / prosecution / target time is accessible that implies the ambiguity are no longer visible and therefore does not impact its related pending task.