Abstract
InfiniCloud 2.0 is World’s first native InfiniBand High Performance Cloud distributed across four continents, spanning Asia, Australia, Europe and North America. The project provides researchers with instant access to computational, storage and network resources distributed around the globe. These resources are then used to build a geographically distributed, virtual supercomputer, complete with globally-accessible parallel file system and job scheduling.This paper describes high level design and the implementation details of InfiniCloud 2.0. A gene sequencing pipeline as well as plasma physics simulation code are used to demonstrate system’s capabilities.
Highlights
The original InfiniCloud system, presented at Supercomputing Frontiers Singapore in March 2015, enabled researchers to quickly and efficiently copy large volumes of data between Singapore and Australia, as well as to process that data using two discrete, native InfiniBand High Performance Clouds [8]
While the unique capabilities of InfiniCloud enabled new ways of processing data, it inspired a whole new range of research questions: Can the entire capacity of the system be aggregated? Do entire data collections need to be copied for processing, or can data be accessed in place? How does the InfiniCloud design scale to an arbitrary number of sites? How we ensure a consistent state of all InfiniCloud clusters? And can the resources across four continents be joined together using the InfiniCortex fabric to create a Galaxy of Supercomputers [14]?
In (Section 1) and (Section 2) we demonstrated the concept, design and implementation of a geographically distributed, High Performance Cloud system, capable of aggregating high performance computing resources available across four continents
Summary
The original InfiniCloud system, presented at Supercomputing Frontiers Singapore in March 2015, enabled researchers to quickly and efficiently copy large volumes of data between Singapore and Australia, as well as to process that data using two discrete, native InfiniBand High Performance Clouds [8]. While the unique capabilities of InfiniCloud enabled new ways of processing data, it inspired a whole new range of research questions: Can the entire capacity of the system be aggregated? In this paper we aim to explore these research questions and propose new ways of utilizing distributed computation, storage and network resources, using a variety of novel tools and techniques. We take advantage of the expansion and enhancement of the InfiniCortex fabric which took place in 2015 [12], which includes full support for InfiniBand subnets and routing, greater available bandwidth and last but not least the growing number of participating sites.
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
