Computing professor designs stream processing system for IoT applications

Smart devices that communicate with other smart devices, known as Internet of Things (IoT) applications, have revolutionized the modern way of life. From self-driving cars to interactive gaming, IoT devices were invented to improve daily life and make simple tasks a thing of the past.

Gartner estimates 20.8 billion connected things to be in use by the end of the year; these applications need a strong digital “architecture” to rapidly process data and to take action upon a user’s immediate command.

Through a National Science Foundation CAREER award of approximately $490,000, Liting Hu—assistant professor in the School of Computing & Information Sciences at the College of Engineering & Computing—is designing and building a stream processing system to benefit time-critical IoT applications and improve their performance. The system will primarily enhance the capabilities of devices involved in factory automation, autonomous vehicles and process automation.

Think ahead to the future, riding in a self-driving vehicle, among several other driverless cars. All these vehicles have numerous sensors installed in them to collect driving activity data. Depending on the role of the sensor, it can measure the distance between cars, advise of incoming traffic, discover a passenger’s favorite music selection and learn the address of their work office.

Hu explains how IoT applications, like self-driving cars, generate a large volume of sensor data. “Under many time-critical scenarios, these data streams must be processed in the blink of an eye to derive actionable intelligence,” Hu says.

With the relative newness of IoT devices, comes many challenges. Hundreds of applications in a confined environment, limited wi-fi connectivity or sensors not being able to hold the same amount of information found in high-memory cloud servers are some of the challenges, to name a few.

Hu and her research team are building what’s referred to as a “Scalable and Adaptive Edge Stream Processing” engine. Stream processing means data that is continuously coming in and being analyzed.

Hu’s CAREER project contains three research parts. First, a dataflow graph abstraction will be implemented, which places the streaming operators to adapt to dynamic network environment. The second part consists of a data shuffling service that customizes the data, while IoT applications are running. Lastly, a fully decentralized architecture will be implemented to process any requests from IoT devices.

The success of the research will improve the performance profiles of a variety of data processing systems, including data analytics systems, mobile data access systems and streaming databases. Once the system is designed, it will be validated by performing real-world experiments.

FIU computer science students will also have an opportunity to dive into this research. Hu teaches an undergraduate course called operating systems and is incorporating a component of the research into a course project on the topics of IoT and big data.

Hu is also developing a new graduate course on stream processing systems; students will learn big data tools in terms of architecture and design.

Through the five-year grant, Hu is collaborating with several Miami-Dade County Public Schools, the nation’s fourth-largest public school district, to donate approximately 100 laptops to Title 1 schools, and offer coding workshops to low-income students.

The workshops will provide students hands-on experience with computer systems, teaching them the fundamentals of coding and computer programming.

“Many times, we see students not pursuing a degree in computer science because they find it intimidating. We want to change that perception,” Hu says. “I love computer science because it is a satisfying feeling to build a computer system from scratch and see it come to life in two to three years. We hope these workshops ignite a passion for STEM in younger generations.”

Hu’s research areas include big data, cloud computing, system virtualization and distributed systems. She obtained her Ph.D. in computer science at the Georgia Institute of Technology. 

The NSF Faculty Early Career Development (CAREER) program supports early-career faculty who have the potential to serve as academic role models in research and education, and lead advances in the mission of their organization.