.png){kind=icon}
.png){kind=icon}
Return
Truthful resource trading for dependent task offloading in heterogeneous edge computing
.png){kind=icon}
PRE
AI
DOI：10.1016/j.future.2022.03.010.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Various Internet of Things (IoT) applications that comprise computation-intensive dependent tasks have been widely popularized. Since these applications have real-time requirements, offloading computation-intensive tasks to edge server devices (ESDs) for expediting the execution of applications has become an important technology. ESDs may incur extra costs when they share computing resources to execute tasks for others. To motivate ESDs to execute IoT applications, this paper designs a reverse auction mechanism that charges the users of applications and rewards the ESDs. Moreover, it can ensure the respective benefits of IoT users and ESDs in resource trading. In the designed auction mechanism, each of the tasks ready for execution in applications counts as an auctioneer, while all ESDs act as the bidder competing for executing the task. Meanwhile, charging IoT users for expediting the execution of applications is formulated as the problem of minimizing the money cost. It is proved to be NP-hard. We propose a greedy strategy for selecting a winning computation resource from the bidder with the polynomial time. Built upon the proposed strategy, all tasks of applications are assigned to the winning ESD with the minimum money cost, and the utility of the winning ESD is maximized as much as possible. Extensive simulations validate that the proposed resource allocation strategy is superior to existing strategies in terms of total money cost, average makespan, and deadline violation rate reduction. (C) 2022 Elsevier B.V. All rights reserved.
Keywords:
Task offloading
Dependent tasks
Reverse auction
Edge computing
Journal
F
IF:
6.1
Papers:
6.8K
Citations:
2.3W