Chameleon: Building an Experimental Instrument for Computer Science as Application of Cloud Computing

Speaker:

Kate Keahey, Mathematics and Computer Science Division, Argonne National Laboratory, Computation Institute, University of Chicago.

Abstract:

Over the last decade, cloud computing has become an established technology, fundamental to all major 21st century economic activities.

However, many open research challenges remain: Is cloud computing a suitable platform for high-performance computing? How can programmable clouds be combined with programmable networks (Software Defined Networking) to offer an end-to-end programmable platform capable of delivering significant qualities of service? How can cloud computing be most effectively leveraged in building an elastic platform for emergent applications such as Smart Cities and other dynamic data stream based instruments? A persistent barrier to developing progress in this area has been the lack of a large-scale experimental platform where they can be explored.

Over the last year, the Chameleon project has built such a platform. The testbed, deployed at the University of Chicago and TACC, consists of almost 15,000 cores, a total of 5 PB of total disk space, and leverages 100 Gbps connection between the sites.

The hardware deployed so far consists primarily of homogenous nodes to support large-scale experiments. On this homogenous base we grafted heterogeneous elements including Infiniband networking, high-bandwidth I/O nodes, storage hierarchy nodes, and GPUs to support a large variety of research projects.

To support Computer Science experiments, ranging from operating system and virtualisation research, Chameleon provides a configuration system giving users full control of the software stack: provisioning of bare metal, reboot, power on/off and console access — but also a fully functioning cloud environment to support educational projects and cloud development.

This talk will describe the goals, the design strategy, and existing and future capabilities of the testbed, as well as some of the research projects in cloud computing our users are working on.

Bio:

Kate Keahey is a computer scientist and a Computation Institute fellow at the University of Chicago. She works within the Argonne National Laboratory's Mathematics and Computer Science Division at the Computation Institute.

She created and leads the Nimbus project, and her research Interests include:

• Virtualisation
• ​Resource management
• Cloud computing.