MRI: Acquisition of the Kentucky Research Informatics Composable Cloud (KyRICC)

Ov e rview: Title: MRI: Acquisition of the Kentucky Research Informatics Composable Cloud (KyRICC) PI: JC Talbert; Co-PIs: VK Bumgardner, HNB Moseley, CM Risko, JN Griffioen Institution: University of Kentucky We propose to purchase and operate a dynamically composable private cloud infrastructure called the Kentucky Research Informatics Composable Cloud (KyRICC). This KyRICC architecture will support complex data analysis pipelines with highly heterogenous hardware needs across individual data analysis steps. Specifically, the KyRICC will integrate 4 subsystems that will enable this dynamically composable cloud infrastructure: 1. A cluster of peripheral-composable compute nodes, allowing for up to 20 GPUs and 12TB of main memory on a single node. Groups of nodes can be dynamically allocated to allow the training and inference of very large DL models and datasets. 2. A next-generation high-speed NVMe- based storage cluster capable of efficiently serving large volumes of data to multi-GPU nodes. Unlike traditional clustered storage system, this composable filesystem allows the partitioning of storage on the project-level, allowing us to isolate data and better manage system performance. 3. A Peta-scale storage system on top of UK’s current research storage system, providing a total of 2.2 PB of storage. 4. Innovative workload management system for dynamic infrastructure composition, workload profiling, model and infrastructure tuning, supporting common pipeline and DL models through templated projects. KyRICC will be managed by UK Advanced Technology Research Computing group composed of UK’s Center for Computational Science (CCS), ITS Research Computing Infrastructure and the Institute for Biomedical Informatics. CCS will oversee policy and procedures. KyRICC resources will be readily accessed by researchers across the state utilizing our latest state-wide high-performance network (Ky Wired), with multiple 100GB/s links from Lexington to Louisville and Cincinnati. We are currently an XSEDE Service Provider and plan to explore the possibility of making this new type of computational resource available to the XSEDE community. Intellectual Merit: Scientific discovery today is driven by computation- and data-intensive research that exploits the enormous amounts of available data. The KyRICC architecture will support massively parallel applications as well as exciting and challenging hardware accelerated and memory-intensive research in big data science that is efficiently executed through dynamic composition of resources. As a result, this project will enable and support a wide range of new research activities, each with its own unique characteristics that are beyond the capacity of our existing HPC infrastructure. Areas of expected breakthroughs in KyRICC-enabled research include deep learning and computer vision; natural language processing and multimodal embedding; computational modeling and simulation with data analytics; and omics analysis and systemic integration. Broader Impacts: KyRICC will provide intuitive access, rapid node customizations, higher bandwidth, and ultrafast storage, facilitating improved algorithm design, software development, and interactive data analysis. KyRICC will accelerate data-driven discovery and computational research education across multiple disciplines by reducing the burden of implementing complex data analysis pipelines. Breakthroughs in KyRICC-enabled research can have important societal benefits in a number of areas: increasing agricultural yields, improving economic competitiveness, and creating new products and markets. KyRICC will be used by hundreds of UK researchers (faculty, staff, and students) and by other computational research collaborators, notably Centre College, Morehead State University. Eastern Kentucky University (EKU), and University of Louisville (UL), Northern Kentucky University (NKU), and Kentucky State University (KSU - a member of Historically Black Colleges and Universities). Thus, the cutting-edge KyRICC and the exciting projects it makes possible will help recruit more students, including students from groups underrepresented in STEM, to computational research and will enhance the research training of many undergraduates, graduate students, and postdocs in Kentucky colleges and universities. The active Kentucky West Virginia NSF LSAMP will provide an existing channel to leverage for outreach and recruiting of interested students from Appalachia. KyRICC will also respond to interest from local industry for education in big data science technologies and tools, thus impacting Kentucky’s and the nation’s economic development.