MONDAY, JULY 19 - FRIDAY, JULY 23, 2010

PORTLAND, OREGON

Jason Stowe (Cycle Computing)
5:20pm Wednesday, 07/21/2010
Cloud Computing
Location: Portland 251

Life science research, including molecular modeling, bioinformatics, proteomics and genomics are ripe with examples of open-source technology. In this presentation, Stowe will present the use of open software in managing state-of-the-art high performance computing (HPC) environments and provisioning auto-configuring software stacks in internal clouds and Amazon EC2.

Research Workflows


Many pharmaceutical and other organizations use pure open-source technology like the Condor Project (http://www.cs.wisc.edu/condor/) and mixed open-source like Sun GridEngine (http://gridengine.sunsource.net/) to manage HPC clusters. The calculations frequently utilize open tools in their software stack, including Gromacs, Blast, and R. Stowe will illustrate the software pipelines commonly used in pharmaceutical research and methods for deploying these software stacks in clouds.

HPC and “Cloudbursting

”
We will discuss using Condor to make the most of internal resources by harvesting compute cycles on dedicated servers, idle workstations and under-utilized virtual environments. Covered architectures will also include features for power management and the ability for researchers to “cloudburst” when internal capacity is exceeded, with specific considerations for synchronizing the data involved for these calculations.