This $505,916 project grant from the National Science Foundation's Computer and Information Science and Engineering program will fund research at the University of Delaware to develop novel performance testing frameworks for serverless cloud computing applications.
The grant aims to address the lack of tools to accurately determine performance for serverless cloud deployments. Over three years, researchers will design correlation-aware statistical techniques including clustered-block bootstrap and regression-based bootstrap to provide three automated testing frameworks. The frameworks will enable unit testing of a single application using intra-application performance correlations from simultaneous invocations. Additional frameworks will integrate testing between two applications and overall performance of serverless workflows by leveraging intra- and inter-application correlations.
Insights from this research seek to deepen understanding of serverless computing performance characteristics and effective analysis using modern statistical methods. Outcomes could help reduce cloud deployment costs and improve user experiences for organizations utilizing serverless platforms, such as in education, research and healthcare.
Generated 1/6/24, 1:52 PM