Energy-efficient offloading of real-time tasks using cloud computing
- PDF / 1,940,768 Bytes
- 16 Pages / 595.276 x 790.866 pts Page_size
- 84 Downloads / 210 Views
(0123456789().,-volV)(0123456789(). ,- volV)
Energy-efficient offloading of real-time tasks using cloud computing Suzanne Elashri1 • Akramul Azim1 Received: 21 April 2019 / Revised: 8 February 2020 / Accepted: 9 March 2020 Ó Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract With the increasing number of sophisticated power-intensive applications, embedded systems require energy-efficient computing devices. Most real-time applications of embedded systems are subject to timing constraints that should be met for the applications to run properly. Unfortunately, local computational devices have limited computation and storage capacity. Moreover, the real-time applications that perform complex computations consume large amounts of power. Therefore, offloading the power-intensive computational tasks to a more powerful entity is an efficient technique to overcome the limited computational resources of local devices and reduces the overall power consumption. In this paper, we propose two algorithms for making an efficient offloading decision for soft and weakly hard (firm) real-time applications while guaranteeing the schedulability of tasks. We have performed different experiments and investigated the technical and economic feasibility of using offloading to perform various processes that require different computational power. Experimental results show that significant power can be saved by offloading the resources of the power-intensive applications into the cloudlet for weakly hard real-time tasks and and cloud for soft real-time tasks. Keywords Cloud Computing Dynamic speed scaling (DSS) Energy-efficient real-time tasks Offloading Supply and demand bound functions
1 Introduction In recent years, we observe a significant development of several software programs that can perform complex computations. In addition, the recent development of capabilities of real-time embedded systems (smart cell phones, ATMs, laptops, broadcasting medical devices, smart thermostat, etc.) has led to a spectacular increase in the power consumption of these systems. Nowadays, realtime embedded devices are capable of performing very complex tasks, which increases their power consumption. For instance, the complexity of applications running on smart devices resulted in a shortening of the battery life. While the quantity and complexity of these applications are increasing on a daily basis, the capacity of lithium-ion batteries, which are the source of power in laptops and smart devices, is slowly improving [1]. Moreover, the amount of heat dissipation is directly proportional to power & Suzanne Elashri [email protected] 1
Ontario Tech University, Oshawa, Canada
consumption. Therefore high rates of power consumption result in high amounts of heat dissipation. Heat dissipation not only deteriorates the performance of computers and embedded systems but also reduces their durability. For instance, a 15 °C increase in temperature from the tolerable temperature can reduce the life of electronic devi
Data Loading...
