.png){kind=icon}
返回
Enabling Collaborative Computing Sustainably Through Computational Latency-Based Pricing
.png){kind=icon}
PRE
AI
DOI：10.1109/TSUSC.2020.2980133.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
Utilizing the idle computing resources from the distributed Internet of Things devices can sustainably increase the computational capacity and thereby effectively alleviate the pressure on resource-constrained devices, which is referred to as collaborative computing. However, extra computing consumption potentially impacts the local computation tasks of collaborative computing devices. Hence, it is essential to design an efficient incentive mechanism for computational resources sharing. Specifically, we consider the collaborative computing system where a user offloads the computation-intensive and latency-sensitive tasks to multiple idle computing devices (ICDs) by a centralized computing sharing platform (CSP). We first propose a computational latency-based pricing mechanism from the perspective of the quality-of-experience performance; then, a game-theoretic computing task allocation approach is developed among the CSP and multiple ICDs to maximize all participants' profit. The CSP first determines the optimal task partition dynamically upon the tasks' arrival; then, the ICDs derive the optimal central processing unit-cycle frequency correspondingly. Simulation results demonstrate that the overall computational latency of our proposed mechanism is significantly decreased, and achieves by at least 13.5 percent improvement compared with the existing schemes. Meanwhile, the profit of all participants is maximum in collaborative computing, which is improved by 41.5 and 27.9 percent for the CSP and the ICDs, respectively.
Keyword:
Collaborative computing
computational latency-based pricing
dynamic task partition
Stackelberg game
期刊
I
IF:
5
论文数:
1.8K
被引数:
4.3K
机构
引用论文
.png){kind=icon}
分享收藏Fog Computing for 5G Tactile Industrial Internet of Things: QoE-Aware Resource Allocation Model
PREAI
分享收藏分享收藏
查看更多内容