Edge Computing for Terminal Query Based on IoT

With the fast development of the IoT, more and more attention has been paid to the acquisition of valuable information from real-time data streams. Nowadays, sampling-based approximate data analytics are properly the most widely used, which trade the output quality for efficiency. What's more, large bandwidth is need for transmitting large amounts of data is these systems. In fact, for some simple queries, we can use the computing ablility of edge nodes to get accurate and lossless results. This idea promotes the design of CalculIot, a real-time data flow system for edge computing and analysis. We construct edge computing nodes and transmission nodes to transfer the computing process originally in the cloud data center to the edge end, making full use of the computing resources of edge nodes, and at the same time, reducing the network transmission pressure and improving efficiency. To showcase the effectiveness of this algorithm, we implemented CalculIot and evaluated its effectiveness using a set of microbenchmarks and real-world case studies. The results illustrated that for terminal query, our algorithm has higher accuracy and efficiency.

I. Introduction

In recent years, with the continuous development of Internet, IoT, big data, cloud computing and other technologies, we have gradually entered a new Internet era of “Everything is connected, and we have comprehensive perception. ” [1]. In order to obtain more data information in a more timely manner to improve productivity, intelligence characteristics and income, a large number of online sensing devices are applied in many fields, such as finance, energy and industry. The IoT is expected to generate 508 ZByte of data by the end of 2019, when billions of new smart sensors and devices will be available. How to deal with such a huge amount of data will be a big challenge for us.