Travel: Workshop on Clusters, Clouds, and Data Analytics for Scientific Computing 2024

University Of Tennessee Knoxville, Knoxville TN

Investigators

Abstract

The research areas of cluster, cloud, and data-analytics computing are undergoing significant transformation due to two major trends. The first trend is the emergence of multicore and hybrid microprocessor designs, which require system designers to prioritize energy usage and application designers to leverage parallelism and data locality. However, the traditional high-performance computing (HPC) software stack is not well-suited for these new architectures with numerous nodes, cores, accelerators, and reduced memory per core. The second trend is the exponential growth in data generated and analyzed by cutting-edge scientific applications and their user communities. This poses challenges not only in processing and managing vast amounts of data but also in making the data accessible to participants in large national and international collaborations, spread across various administrative domains and utilizing different resources like clusters, clouds, and data analysis tools. These new conditions present several complex design and deployment challenges for the cyberinfrastructure research community. They must address issues related to scalability, programmability, performance, interoperability, resilience, resource virtualization, data logistics, and system management. The Workshop on Clusters, Clouds, and Data Analytics in Scientific Computing (CCDASC) will be held to tackle these challenges and facilitate fruitful discussions. This workshop aims to bring together leading-edge researchers to collaboratively review, analyze, and gain new insights into these developments, and translate these advancements into significant benefits for the scientific community.   The workshop aims to achieve its goal by focusing on the commonalities and interactions between clusters, clouds, and data analytics and how they can support data and compute-intensive collaborations. These computational infrastructures share architectural similarities, and addressing issues related to cloud computing encompasses those concerning clusters. The growing importance of data-driven research and cross-domain collaboration underscores the need for solutions to data interoperability and logistics challenges. Key topics to be explored in the workshop include overcoming scalability challenges for high performance and computational efficiency, developing portable software for various platforms (clusters, grids, clouds), creating common protocols and APIs for seamless data movement and analysis across resource domains, designing easier programming paradigms for parallel systems, and solving system management problems arising from multicore architectures, resource virtualization, and data-intensive applications. The workshop format fosters synergistic effects among researchers from different focus areas, facilitating coordination for high-impact research that spans multiple domains. The workshop's tasks include surveying and analyzing deployment, operational, and usage issues for clusters, clouds, and data analytics, documenting the current state-of-the-art in each area, sharing experiences of research communities and science domains benefiting from these technologies, exploring interoperability among clouds and grids, and identifying future research directions in light of disruptive trends and technological gaps. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.