Edge computing has evolved to be a promising avenue to enhance the systemcomputing capability by offloading processing tasks from the cloud to edgedevices. In this paper, we propose a multi- layer edge computing frameworkcalled EdgeFlow. In this framework, different nodes ranging from edge devicesto cloud data centers are categorized into corresponding layers and cooperatetogether for data processing. With the help of EdgeFlow, one can balance thetrade-off between computing and communication capability so that the tasks areassigned to each layer optimally. At the same time, resources are carefullyallocated throughout the whole network to mitigate performance fluctuation. Theproposed open-source data flow processing framework is implemented on aplatform that can emulate various computing nodes in multiple layers andcorresponding network connections. Evaluated on a mobile sensing scenario,EdgeFlow can significantly reduce task finish time and is more tolerant torun-time variation, compared to traditional cloud computing and the pure edgecomputing approach. Potential applications of EdgeFlow, including networkfunction visualization, Internet of Things, and vehicular networks, are alsodiscussed in the end of this work.
展开▼
