Recently Internet of Everything has emerged as integration of various IoT machines to which cloud computing provide storage to data and processing power to the geographically distributed IoT devices. In order to boost the efficiency of the system and Quality of Service(QoS), fog layer is added in the existing cloud infrastructure. Due to continuous increase in the time sensitive applications, reduction in latency is a crucial issue in the fog computing paradigm. Therefore, the main objective of this work is to reduce the latency in fog computing. To achieve this objective, popularity based caching is performed in this work by majorly focusing on the interest of the users. In this context, first clustering of the IoT devices is performed on the basis of their interests and distance between them using spectral clustering technique and then each cluster is mapped with the fog node such that caching of the popular files is done effectively. To further reduce the latency, in case of cache miss, the Device to Device (D2D) communication is used. Finally, association rules are also used to predict the future demands of the IoT devices. Performance analysis of the proposed scheme shows that the proposed method outperforms the other existing caching methods.

近年来，万物互联已经成为各种物联网机器的集成，云计算向这些物联网机器提供了数据存储和对地理分布的物联网设备的处理能力。为了提高系统效率和服务质量（QoS），在现有云基础架构中添加了雾层。由于时间敏感型应用程序的不断增加，延迟的减少是雾计算范例中的关键问题。因此，这项工作的主要目的是减少雾计算的延迟。为了实现该目的，主要通过关注用户的兴趣来进行基于流行度的缓存。在这种情况下，首先，使用频谱聚类技术根据物联网设备的兴趣和它们之间的距离对物联网设备进行聚类，然后将每个聚类与雾节点进行映射，从而有效地完成流行文件的缓存。为了进一步减少延迟，如果发生高速缓存未命中，则使用设备到设备（D2D）通信。最后，关联规则还用于预测物联网设备的未来需求。对该方案的性能分析表明，该方案优于其他现有的缓存方法。