High concurrency massive data collection algorithm for IoMT applications

The symbiotic development of machine learning (ML) and artificial intelligence (AI) is amplifying the value of the Internet of Medical Things (IoMT). Doctors are able to reach actionable conclusions faster and more reliably when dealing with large volumes of streaming data from networked medical devices. However the Internet of Things (IoT) or sensor network has a large number of base stations transmitting data to the data center server, the data center server will face challenges in collecting, parsing, and processing data. Based on the existing technical solutions, when the number of wireless sensor network base stations is large, the data collection of the IoMT system will have a concurrent bottleneck, which will cause the data collection failure and have a catastrophic impact on the application of the IoMT. This paper proposes a highly concurrent and massive data collect algorithm for IoMT applications. This algorithm uses the principle of separation of reception and processing, distributed parallel processing and multi-threading technology, and combines the highly concurrent data transmission channel provided by TCP/IP to provide a set of independent data receiving components. This component receives the data of the IoT base station, and then simply processes the data and puts it into the distributed message system to complete the sensor data receiving function. It also provides a data processing cluster. Each node of the cluster starts multiple data processing unit, each data processing unit separately obtains sensor data from the distributed message system, processes the data, and delivers the processing results to the application. The experimental results show that the algorithm proposed in this paper has a high ability of parallel collection of IoMT data.