Collaborative Research: Chameleon Phase III: A Large-Scale, Reconfigurable Experimental Environment for Cloud Research

点击翻译按钮获取中文摘要

Chameleon is a deeply reconfigurable experimental testbed supporting Computer Science (CS) systems experimentation. The platform consists of two sites, University of Chicago (UC) and Texas Advanced Computing Center (TACC), along with support functionalities from Northwestern University and the University of North Carolina. the platform balances investment in large-scale hardware to support Big Compute and Big Data experimentation with diversity reflected by smaller clusters of graphic processing units (GPUs), field programmable gate arrays (FPGAs), specialized architectures, and innovative networking hardware. Users can reconfigure this hardware at bare-metal level, boot from custom kernel, get access to serial console, or provision and reconfigure Software Defined Networking (SDN)-enabled switches. Since its public availability in July 2015, Chameleon has attracted a community of over 4,000 users working on over 600 education and research projects.In Phase III, Chameleon will greatly expand its core capabilities by adding new hardware and new features. The range of supported experiments will expand to include Internet of Things (IoT) and edge computing, a broader range of networking experiments, and experimentation with disaggregated hardware. The system will also expand its support for reproducibility by providing tools that allow investigators to package their experiments for replication, and making it possible to publish, discover, and cite them easily. Finally, phase 3 will see a continuation of work on packaging CHameleon Infrastructure, known as CHI-in-a-Box, which allows others to deploy and manage testbeds similar to Chameleon. The Chameleon project will continue its mission of providing a platform for CS research and education in cloud computing, a critical skillset for future computing professionals. Unlike traditional experimental platforms, Chameleon is built on a mainstream open source cloud technology (OpenStack); this not only means that its users and operators gain familiarity with mainstream cloud functions simply by working with the testbed – but also that open source contributions by the Chameleon project directly impact a system widely used in the cloud industry. Finally, through its support of experiment management and repeatability, Chameleon impacts CS experimental methodology and makes more efficient research possible. To use Chameleon or learn more about the system, visit www.chameleoncloud.org. For anyone interested in the development side of the system or the packaging of Chameleon, i.e., CHI-in-a-Box, visit the github repository at https://github.com/ChameleonCloud/chi-in-a-box/wiki. If a user would like to explore traces from Chameleon to date, see https://www.scienceclouds.org/cloud-traces/. These sites will be maintained throughout the duration of the project.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.