Information Technology Reference
In-Depth Information
around the globe and for whom the 'clouds of Magellan' - two small galaxies in the southern
sky - were named. Pete Beckman, director of Argonne's Leadership Computing Facility and
project lead, said “We know that the model works well for business applications, and we are
working to make it equally effective for science.” (Argonne National Laboratory, 2009a).
Another interesting cloud-based HPC solution developed by the US DOE at its Argonne
National Laboratory
3
is called 'Nimbus' (a Latin word for 'cloud' or 'rain storm'). Nimbus is
an open source cloud software infrastructure that allows users to deploy clusters of appliances
(i.e., VMs) on remote resources, similar to Amazon's EC2 (see Argonne National Laboratory,
2009b; Keahey, Tsugawa, Matsunaga and Fortes, 2009). The advantages of using cloud
computing for scientific purposes were illustrated recently by researchers working on the
STAR nuclear physics experiment at Brookhaven National Laboratory's Relativistic Heavy-
Ion Collider. The researchers needed simulation results to present at the Quark Matter physics
conference but found it difficult to do so as all the computational resources at their disposal
were either committed to other tasks or did not support the environment needed for STAR
computations. However, the STAR researchers were finally able to use the Nimbus platform
in order to dynamically provision virtual clusters on commercial cloud computers and run the
additional computations.
Jerome Lauret, software and computing project leader for the STAR project, commented:
'The benefits of virtualization were clear to us early on....We can configure the virtual
machine image exactly to our needs and have a fully validated experimental software stack
ready for use. The image can then be overlaid on top of remote resources using infrastructure
such as Nimbus'. (Argonne National Laboratory, 2009b)
He added that with cloud computing 'a 100-node STAR cluster can be online in minutes
[whereas] Grid resources available at sites not expressly dedicated to STAR can take months
to configure' (Ibid.)
The STAR scientists initially developed and deployed their VMs on a small Nimbus
cloud configured at the University of Chicago. Then they used the Nimbus Context Broker
4
to configure the customized cloud into grid clusters which served as a platform for remote job
submissions using existing grid tools. However, these resources soon proved insufficient to
support STAR production runs.
In order to meet these needs, the Argonne Nimbus team turned to Amazon EC2 for help.
The team developed a Nimbus gateway which allowed them to move easily between the small
Nimbus cloud and Amazon EC2.
Kate Keahey, the lead member of the Nimbus project, commented:
'In the early days, the gateway served as a protocol adapter as well…But eventually we
found it easier to simply adapt Nimbus to be protocol-interoperable with EC2 so that the
scientists could move their virtual machines between the University of Chicago cloud and
Amazon easily' (Ibid.)
3
The Argonne National Laboratory is operated by the University of Chicago for the United States Department of
Energy (DOE).
4
This Context Broker is a service that allows users to coordinate large virtual cluster launches automatically and
repeatedly. It is used to deploy 'one-click' virtual clusters that function right after launch as opposed to
launching a set of unconnected' VMs which most VM-on-demand services provide. It also provides a facility
to 'personalize' VMs (e.g., enable access policies).
