Intelligent devices that can communicate with other functional devices are called Internet of things (IOT) devices, which have revolutionized the modern lifestyle. IOT devices can improve every aspect of daily life, from self driving cars to interactive games. By the end of this year, Gartner has estimated that 20.8 billion companies will be able to conduct research on this kind of data architecture and take quick action based on the data architecture of such products.
According to foreign media reports, Li Ting Hu, an assistant professor at the school of engineering and information science at Florida International University, is designing and building a stream processing system to benefit time sensitive IOT applications and improve their performance. The system can enhance the functions of factory automation, autopilot and process automation equipment.
In the future, autopilot will install a large number of sensors to collect driving activity data. Depending on the roles of different sensors, it can measure the distance between vehicles, provide instructions for the upcoming traffic situation, find out the love of the passengers’ favorite music and simultaneous interpreting the address of the passenger office. Autopilot is also a IoT application, which generates a large amount of sensor data. In many time critical situations, such data streams must be processed instantaneously in order to obtain operational intelligence.
With the continuous updating of IOT devices, there are also many challenges. For example, there are hundreds of applications in a limited environment, and the Wi Fi networking function is limited, or the sensor can not hold the same amount of information as the high memory cloud server.
Therefore, Hu and his research team are building an “extensible and adaptive edge stream processing” engine, which refers to continuously inputting and analyzing data. The research includes three parts: first, the abstract graph of data flow is executed to make the flow operator adapt to the dynamic network environment; secondly, when the IOT device is running, it implements the data transfer service that can customize the data; finally, it implements a completely decentralized architecture to handle any request of the IOT device.
If the research is successful, it will improve the performance profile of various data processing systems, including data analysis system, mobile data access system and streaming database. Once the system is designed, it can be verified by real experiments.
Responsible editor: Tzh