High Performance Computing (HPC) Data Center Planning and TCO: A Case Study and Roadmap
HPC Center Planning and Operations
TimeSunday, 17 November 20191:30pm - 5pm
DescriptionBuilding upon the “Data Center Planning and Design” Tutorials presented at SC18, SC17 and SC16 and attendees’ feedback, this tutorial provides best practices and lessons learned gleaned from multiple HPC data center infrastructure capacity planning, total cost of ownership (TCO) analysis and business case justification initiatives over the last 15 years and, significantly, provides a deep-dive case study and roadmap around Lawrence Livermore National Laboratory’s (LLNL) methodology to accommodate future HPC systems. The presenters understand that the HPC data center mechanical, electrical and structural infrastructure can be a significant enabler or roadblock to the timely deployment and optimized performance of HPC systems. We are encouraged to see more TCO/business case dialogue in the HPC community than in years past and will share firsthand experience about how participants can develop/enhance their organizations’ HPC facility infrastructure to match their deployment methodology. Topics covered include identifying and managing stakeholders, defining the mission, developing a facilities infrastructure roadmap, performing HPC vendor surveys, developing program requirements and key performance parameters, following project processes, developing alternative studies, and establishing scope and budget expectations. This tutorial will improve your technical knowledge, project management skills, and financial confidence to navigate this process and drive successful outcomes.