Cloud HPC Demystified: Best Practices of Executing and Managing Simulation Workloads on the Cloud

This presentation was made at CAASE18, The Conference on Advancing Analysis & Simulation in Engineering. CAASE18 brought together the leading visionaries, developers, and practitioners of CAE-related technologies in an open forum, to share experiences, discuss relevant trends, discover common themes, and explore future issues.

Resource Abstract

High performance computing on the cloud offers a variety of advantages to simulation engineers and teams, notably instantly scalable computing architecture, broad software access, and integrated automation tools. However, transitioning these simulation workflows directly to the cloud can be exceedingly complex to first time users and their IT organizations. They can be difficult to monitor, manage and control-- and more importantly they can be challenging to perform securely. This 90 minute workshop will cover how to easily overcome these challenges and convey other cloud fundamental concepts (beyond just bursting) in a short presentation.

Some common questions engineers may face when considering cloud computing are “Which of my workflows are most suitable for cloud computing?”, “How do I address file size or file count issues with the cloud?”, and "How do I connect to my license server or manage on-demand license models?" The presentation portion will address identifying workflow suitability, the best practices for transitioning to the cloud, and then conclude with a live end-to-end workflow on the Rescale ScaleX platform.

Following the workflow demonstration example, there will be a short discussion session in small groups where key barriers and questions can be addressed and shared with the rest of the participants.

In addition to addressing the transition of simulation workflows to the cloud, attendees should also exit the workshop with a grasp of how overall data management is treated in the cloud. This includes strategies for uploading and retrieving large files and data, sets how data is stored and organized between different cloud data mechanisms, and the relevant security measures for data in transit and at rest.

The workshop will also cover touch on relevant cloud administration functions, such as how job monitoring is handled, how jobs are shared within teams, and how software and on-demand licensing can be supported.