Microservices and Data-Sharing Cyberinfrastructure for Supporting Collaborative

Research and Education

Over the Internet and current cloud platforms, there are multiple and valuable resources for researchers and academia. These resources include data and codes for different tasks in diverse fields. In research, many algorithms and solutions are open-source, and those resources can facilitate the researcher’s path to a successful project.

Even private clouds, like Amazon Web Services (AWS) and Microsoft Azure, offer specialized algorithms for multiple computational areas. The problem is that finding, executing, and reproducing these resources is no easy task. Researchers have to deal with platform/library-dependent codes and lack of data. Although multiple approaches have been proposed to share codes (e.g. GitHub) and data (e.g ResearchCompendia), there is a lack of an infrastructure to find and execute platform-library independent codes as services, using proprietary or shared data for development and comparison purposes with the idea to speed up the research process.

Therefore, this project proposes a MIcroservices and DAta-Sharing cyberinfrastructure, called MIDAS-Cloud, that provides platform-independent microservices to support collaborative research and enhance education. MIDAS-Cloud is composed of three collaborative approaches.

First, MIDAS-Cloud will provide microservices for researchers to execute and evaluate code in the cloud with their own or preexisting data using an API mechanism. The researcher can use and control the microservice without being aware of the platform and libraries’ specifications. For example, suppose that researcher A wants to calculate the blood pressure measurements from the data retrieved by wearable devices; and suppose that MIDAS-Cloud contains a microservice that allows the blood pressure estimation using an algorithm designed by another researcher B; researcher A can execute the algorithm by sending API requests and his own (or pre-existing) data to the cloud and retrieve the output. This approach is called microservices-consumer.

The second approach is called microservices-sharing. MIDAS-Cloud will count on a mechanism that allows researchers and developers to upload microservices and API specifications for certain algorithms. This approach will enrich the pool of microservices available in the infrastructure.

The third approach is called data-sharing-mechanism. MIDAS-Cloud will enable a data repository from multiple areas that allow researchers to have preliminary data for their research. For example, suppose research A is looking for data that helps him to test a machine learning algorithm for sleep apnea prediction; MIDASCloud may have an open database with sleep data that can be useful for that specific research. Furthermore, MIDAS-Cloud can enrich the academy by providing students in different courses with code and data to test the course’s objectives.

The impact of having such infrastructure is tremendous as the microservices can be used even in IoT and embedded systems, as long as the device follows the API requirements.

Publications

  1. Enabling Cyberanalytics using IoT Clusters and Containers. Traore, Soin; Valero, Maria; Shahriar, Hossain; Zhao, Liang; Lee, Ahyoung. IEEE STPSA 2022: Security, Trust, and Privacy for Software Applications, 2022.
  2. Secure Cloud-based IoT Water Quality Gathering for Analysis and Visualization. Traore, Soin; Valero, Maria; Gruss, Amy. 2022 KSU CONFERENCE ON CYBERSECURITY EDUCATION, RESEARCH AND PRACTICE, 2022.