PROJECT TITLE :

Scheduling Real-Time Parallel Applications in Cloud to Minimize Energy Consumption

ABSTRACT:

The concept of cloud computing has emerged as an important paradigm in recent years. Cloud computing enables users to remotely process their applications by providing them with scalable resources such as CPU, memory, disk, and IO devices. Consumption of electricity is a significant contributor to the overall cost of running a cloud computing platform. Therefore, the purpose of this article is to present a scheduling algorithm that is both energy efficient and capable of processing a user application that has a real-time requirement. This issue is modeled as a non-linear mixed integer programming problem to facilitate its analysis. To begin, we provide an optimal closed-form solution to its relaxation problem. This solution's overarching goal is to minimize the amount of energy that is consumed, and it does not take into account any real-time requirements. We propose a method for adjusting the placement of tasks and the allocation of resources in order to meet real-time requirements. This method achieves a satisfactory balance between the amount of energy consumed and the amount of time required to complete tasks. After the placement of the tasks has been finalized, we find two optimal resource allocation strategies that are equivalent to one another. The next step that we propose taking is modifying the start time of the task execution so that the amount of time it takes to finish an application can be reduced even further. Our proposed method finds a schedule that, on average, uses 30 and 20 percent less energy than enhancement heterogeneous earliest finish time (E-HEFT) and genetic algorithm, respectively. These findings were demonstrated by experimental findings on two real-case benchmarks and extensive synthetic applications. In addition, the proposed method has a higher success rate in finding a schedule that is feasible than the other methods, and its computation time is comparable to that of E-HEFT, but significantly lower than that of the genetic algorithm.