Authors: Heather Savoy (USDA-ARS SCINet Program), Brian Stucky (USDA-ARS SCINet Program), Moe Richert (USDA-ARS SCINet Program), Haitao Huang (USDA-ARS SCINet Program), Kyle Chard (University of Chicago, Argonne National Laboratory (ANL)), Frank Wuerthwein (San Diego Supercomputer Center (SDSC), UC San Diego), Sean Cleveland (University of Hawai'i, System Information Technology Services - Cyberinfrastructure), Sergiu Sanielevici (Pittsburgh Supercomputing Center (PSC)), Stephen Chan (Lawrence Berkeley National Laboratory (LBNL)), Preston Smith (Purdue University Rosen Center for Advanced Computing)
Abstract: This BoF session will address user experience challenges that arise from geographically dispersed computing resources, such as when an organization operates multiple HPC clusters or wishes to combine on-premises and cloud-based compute services. A series of speakers will provide an overview of current perspectives on and solutions for making dispersed computing resources available to user communities. We invite participants to engage in a facilitated follow-up discussion to identify key unresolved hurdles and document emerging community best practices for providing the best possible user experience in geographically dispersed HPC settings.
Long Description: HPC users increasingly have a multitude of geographically dispersed computing resources at their disposal, which can include multiple, non-colocated HPC clusters, cloud computing services (potentially from multiple vendors), and local computing hardware. Although this offers great flexibility, it also introduces significant challenges for both users and system designers and operators, including: data storage and movement between computing resources; providing a consistent user experience across computing systems; and providing a single, unified “point of entry” for launching computing jobs. The goals of this BoF session are to: 1) survey the current landscape of solutions for addressing these challenges; 2) identify key unresolved hurdles; and 3) document emerging community best practices for providing the best possible user experience in geographically dispersed HPC settings.
We will accomplish these goals through a series of brief presentations and guided discussions with session participants. The session presentations will address our first goal by providing an overview of currently available solutions. These presentations will feature leading experts and innovators in this problem space from multiple national labs, universities, and NSF-funded HPC initiatives. Our second and third goals will be addressed via carefully designed, structured participant discussions and question answering. Two of the session organizers have specific training in facilitating community discussions. After the conference, we will develop a brief report summarizing session outcomes and suggesting possible directions for future work on this topic.
To the best of our knowledge, this topic has not been directly addressed by any recent BoF sessions, but we are confident that the questions addressed by this BoF are of increasing importance for many HPC users and operators. For example, even if an organization does not manage multiple on-premises HPC systems, they may wish to augment their on-premises capabilities with cloud-hosted infrastructure. Similarly, although HPC users can benefit from access to an array of computing solutions, navigating a dispersed computing environment remains challenging. Thus, we anticipate that this BoF session will be of interest to conference attendees from a variety of backgrounds and experiences.
Back to Birds of a Feather Archive Listing