Prev: P5.10 Next: P5.13

P5.11: Major, Brian
Brian Major (National Research Council Canada)
JJ Kavelaars (National Research Council Canada)
Sébastien Fabbro (National Research Council Canada)
Daniel Durand (National Research Council Canada)




Theme: Science Platforms: Tools for Data Discovery and Analysis from Different Angles
Title: Arcade: A User Focussed, Visual Compute Environment in CANFAR

For a number of years CANFAR (Canadian Advanced Network for Astronomy Research) has offered virtual machines (VMs) as a way to do both interactive computing and batch processing in the cloud. A VM is a general and flexible base offering that can suit nearly any given astronomy compute project, but demands of users a thorough understanding of the intricacies of software installation and maintenance, and requires significant effort to achieve initial benefits. Arcade is an effort by the Canadian Astronomy Data Centre (CADC) to offer the CANFAR community an astronomy-focussed, easier-to-use, and more intuitive science platform. Arcade is composed of a number of pre-built, specialized application bundles ('game consoles') that appear in a graphical desktop environment. They run independently of each other in the cloud, allowing Arcade to optimize console execution to their particular resource requirements. Consoles have user-specific access to a shared file system and other CANFAR cloud services, outfitting users with a variety of computing, analysis and storage tools. At the CADC we see a lot of potential in Arcade and will be planning its evolution based on a number of key questions. How can scalability be best achieved? How can we reduce the burden of software and infrastructure maintenance for users and operators? Can we allow users to customize their Arcade experience? How can consoles be launched in batch processing? How can we best leverage open source technology and development from other projects? We shall discuss the core concepts of Arcade and explore its potential in respect to these questions and from feedback from the astronomy computing community.

Link to PDF (may not be available yet): P5-11.pdf