IoT edge computing and deep learning analytics: A survey

The group of real-world physical devices like sensors, machines, vehicles and various things connected to Internet is called as Internet of things (IoT). The major challenge in IoT is that it is fully dependent on the cloud for all kinds of computation and some of the IoT based applications such as Smart Navigation system, Smart Health monitoring system needs new requirements such as mobility, location-based awareness etc. cannot be fulfilled in cloud processing. ICT’s three pillars namely computing, network and storage faces some challenges in terms of processing and structuring the data while using formal Cloud computing methods. Hence Edge Computing arrives with the processing and storage in the edge of the networks which is very close to data sources when compared to Cloud Processing. In other dimensions, Deep learning is a potential method for gleaning accurate and usable information from the unprocessed IoT sensor data. Therefore, in this detailed review work, we first address the latency issues, security and privacy issues associated with edge computing, then we introduce IoT deep learning for IoT data analysis in Edge Computing devices. Some of the objectives of Edge computing in Service level are Latency minimization, Network Management, Cost Optimization, Data Management, Energy Management, and Resource Management. The review work still remains absent in making a thorough review on the recent advancements of data security analysis in edge computing. Furthermore, we focus on the advantages and disadvantages of the existing works on security and data processing in edge computing environment and highlighting the open issues.