SI2011 Scalable Scientific Applications and Workflows using Cloud Services
The track is geared towards scientific software developers who wish to make their applications accessible to different types of clients and workflow environment, and scalable in different types of computing environment, GPU, cluster, or cloud resources. Users will learn how to use various types of computing environments through direct logins or through Web services coupled to scalable back-end infrastructures. The computing infrastructure covered include Rocks cluster environment, TeraGrid, and the Amazon EC2 based cloud computing environment. The Web services interface and portal environment to the computing infrastructure will be illustrated using the Opal toolkit, which exposes scientific applications as Web services to visualization tools and workflow environments. Advanced topics such as Metascheduling, Grid monitoring and workflow tools will also be introduced.
Each session includes lectures and some hands on exercises.
Projected Outcome By the end of the sessions, the developers should be able to understand the types of computing environments available, and how to deploy scientific applications with the scalability required by end users. They will be able to deploy applications as Opal-based Web services, and provide basic Web interfaces or portal environments for accessing these application level services in a transparent fashion. They will also be able to develop workflows using NBCR CADD pipeline for customized use.
Instructors: Jane Ren, Nadya Williams, Sriram Krishnan, Ph.D., Raj Singh, Ph.D., Michel Sanner, Ph.D., Wilfred Li, Ph.D.
Time: M-Th, 2:00 pm to 5:30 pm, Aug 1-4, 2011
Location: Atkinson Hall, Room 3004
Prerequisites: Familiarity with Unix environment is helpful but NOT required.
- Monday: Software as a Service and the Opal Toolkit
- Instructors: Nadya Williams, M.S., Jane Ren
- Students will learn how to take advantage of the cluster and cloud computing environment using the Opal toolkit.
- Talk slides:
- Tuesday: Application Deployment Case Studies
- Instructors: Jane Ren, Nadya Williams
- Learn the basics of Rocks and Rocks rolls for automated application deployment and reproducible computing environment. Deploy your own application on Amazon EC2 as an Opal Web Service
- Session hands on materials:
- Talk slides
- Wednesday, Aug 3: Scientific Workflow Development using Opal Services and Vision
- Instructors: Michel Sanner, Ph.D., Jane Ren, Wilfred Li, Ph.D.
- Students will learn how the NBCR CADD pipeline is developed using Vision and Opal Services, and learn to integrate their own application into the CADD/Vision environment. The Opal plugins for Kepler, Vistrails, and Taverna will also be demonstrated.
- Talk slides: workflows
- Thursday, Aug 4: Computing Trends and Application Development
- Instructors: Raj Singh, Ph.D., Sriram Krishnan, Ph.D.
- Students will learn about popular programming techniques such as CUDA programming, and MAP-Reduce programming model. A general overview of current computing environment available as campus wide or nation wide computing infrastructure will also be given.
- Introduction to GPU Computing: CUDA programming and case studies using TxBR
- Introduction to Map-Reduce and Hadoop