Due to the big data exchange on the Internet of Things, proper routing and selecting the best routes for fast data transmission improve network performance. There are major challenges, like high delay, when cloud computing is used. Therefore, one solution is to use other schemes, such as fog computing. In fog computing, all data is not sent to the cloud and the fog nodes close to objects are used for data processing. This reduces the network delay. In this paper, we propose an overlapping clustering method called MFCT-IoT to select the best cluster head nodes to guarantee the fast data transfer from objects to fog nodes. The selected cluster head nodes are responsible for sending the collected data to the closest fog nodes in the network edge. Upon receiving the data, the fog nodes process it, and if a response is ready, they respond immediately to the object. Otherwise, they merge and transmit the data to the cloud servers, which are considered as the root node of the proposed hierarchical tree. After processing, the merged data is sent to the object. We compare the proposed scheme with two schemes, including ERGID and EECRP. These schemes are evaluated based on various criteria, including the response time, packet delivery ratio, end-to-end delay, network lifetime, and energy consumption. The results indicate that the proposed method outperforms others in terms of all criteria.

翻译：由于物联网中大数据交换的需求，合理路由和选择最佳传输路径可提升网络性能。当采用云计算时，存在高延迟等重大挑战。因此，一种解决方案是采用其他方案，例如雾计算。在雾计算中，所有数据无需全部传输至云端，而是利用靠近物体的雾节点进行数据处理，从而降低网络延迟。本文提出一种名为MFCT-IoT的重叠聚类方法，用于选择最优簇头节点，确保数据从物体到雾节点的快速传输。被选中的簇头节点负责将收集的数据发送至网络边缘最近的雾节点。雾节点接收数据后进行处理，若可立即响应则直接向物体回复；否则，将数据合并后传输至云服务器（即所提层次化树的根节点）。经过处理后，合并数据被发送至物体。我们将所提方案与ERGID和EECRP两种方案进行对比，基于响应时间、数据包传输率、端到端延迟、网络生存周期和能耗等多个指标进行评估。结果表明，所提方法在所有指标上均优于其他方案。