In this emerging world of connected devices, the need for more computing devices with a focus on delay-sensitive application is critical. In this paper, we propose a priority-queue based Fog computing architecture combined with dynamic scalability of fog devices, which not only reduces the delay experienced by delay-sensitive tasks by categorizing the delay-sensitive and delay-insensitive tasks, but also dynamically allocates the fog devices within the network depending upon the computation load for reducing the power consumption. The results show that the proposed algorithm is able to achieve a significantly lower delay for both delay-sensitive and -insensitive tasks when compared with other related schemes with a 14.5% lower power consumption.

翻译：在这个连接装置的新兴世界中,需要更多计算机设备,重点是延迟敏感应用。 在本文中,我们建议建立一个基于优先排的烟雾计算结构,加上雾装置的动态可缩放性,这不仅通过对延迟敏感和延迟不敏感的任务进行分类来减少延迟敏感任务造成的延迟,而且根据减少电耗的计算负荷,动态分配网络中的雾装置。 结果表明,与低电耗14.5%的其他相关计划相比,拟议的算法能够大大降低延迟敏感和不敏感的任务的延迟。