Optimizing Edge Node Efficiency Through Data Reduction Using Dimension Reduction Techniques

Cloud computing is a remote state-of-the-art method that facilitates data storage and analytics. It does not guarantee a high processing speed due to its limitations. Fast computation can be achieved with edge computing, which works very efficiently in low-latency scenarios. Latency is a crucial factor in today's networking devices. Edge node latency is discussed in association with the time complexity of different dimension reductions. This paper consists of the various techniques of dimension reduction for data optimization. This paper illustrates the effects of data size on edge node latency. Dimension-reduction methods are used to process data, for better performance. Data sizes affect latency in cloud, fog, and edge processing, as well as various techniques for dimension reduction processing, latency versus data size, and energy consumption versus data size. The paper proposes using edge as compared to the cloud for all low latency devices like health care self-driving cars industrial IoT and smart cities. A fast algorithm can be achieved by comparing the time complexity of various dimension reduction techniques. All simulations are conducted using Google Collaboration on a Python 3 Google Compute Engine backend.