The Advantage of Computation Offloading in Multi-Access Edge Computing

Abstract

Computation offloading plays a critical role in reducing task completion time for mobile devices. The advantages of computation offloading to cloud resources in Mobile Cloud Computing have been widely considered. In this paper, we have investigated different scenarios for offloading to less distant Multi-Access Edge Computing (MEC) servers for multiple users with a range of mobile devices and computational tasks. We present detailed simulation data for how offloading can be beneficial in a MEC network with varying quantitative mobile user demand, heterogeneity in mobile device on-board and MEC processor speeds, computational task complexity, communication speeds, link access delays and mobile device user numbers. Unlike previous work where simulations considered only limited communication speeds for offloading, we have extended the range of link speeds and included two types of communication delay. We find that more computationally complex applications are offloaded preferentially (especially with the higher server:mobile device processor speed ratios) while low link speeds and any delays caused by network delays or excessive user numbers degrade any advantages in reduced task completion times offered by offloading. Additionally, significant savings in energy usage by mobile devices are guaranteed except at very low link speeds.

Edge AI: A survey

“…With the emergence of IoT technology and high mobile-intensive applications such as 4 K surveillance cameras, virtual reality and gaming applications etc., requirements for high processing power have increased. However, due to hardware limitations of wireless devices' computation, offloading must be performed to increase functionality, but overloaded offloads can degrade performance [85]. The solution is an edge server: a wireless device can transfer a series of tasks to the edge server for offloading.…”

To Offload or Not? An Analysis of Big Data Offloading Strategies from Edge to Cloud

“…Using this process, bits are converted into numbers of instructions [15]. With chosen values of instructions per second as computing speeds [15], the computation times for tasks can be calculated [15]- [17]. In general, the smaller the bpi value of an application, the more computationally complex is the task.…”

Identification of the Key Parameters for Computational Offloading in Multi-Access Edge Computing

“…However, in such studies, little attention has been paid to the importance of various parameters involved in decisionmaking in computational offloading problems. Previous work from the authors investigated the impact of mobile devices' computational power (MDs) and MEC servers on job completion times [10]. The work did not take into account CPU workloads and the link speed that connect MD to MEC servers.…”