Load Balancing and Grid Computing

1. Load Balancing and Grid Computing David Finkel Computer Science Department Worcester Polytechnic Institute

2. References • “The Anatomy of the Grid”, Ian Foster, Carl Kesselman, Steven Tuccke, International Journal of Supercomputer Applications, 2001 • “A Performance Oriented Migration Framework for the Grid”, Satish S. Vadhiyar and Jack J. Dongarra, Proceedings of CCGrid 2003, Third IEEE/ACM International Symposium on Cluster Computing and the Grid • Innumerable papers by PEDS members Finkel, Wills and Finkel, and Claypool and Finkel, with additional co-authors.

3. What is the Grid? (Foster et al paper) • Distributed computing infrastructure for advanced science and engineering • Runs over the Internet, potentially world-wide • Several approaches have emerged: Paper discusses Globus Toolkit

4. The Grid Concept • Coordinated resource sharing and problem solving in dynamic, multi-institutional virtual organizations. • Highly controlled, with resource providers and consumers defining what is shared and the conditions of sharing. • Issues to address: Protocols, privacy, security, costs, …

5. Related approaches • Application Service Providers • Storage Service Providers • CORBA • DCE • Volunteer Computing (SETI @ home, Distriblets, SLINC)

7. Fabric Layer • Provides access and control to resources • Resources: Computational, storage, network • Enquiry functions: to determine characteristics and state of a resource • Management functions: Start, stop computations, reserve bandwidth

8. Collective Layer • Protocols and services not associated with a particular resource • Directory services for discovery of resources • Co-allocation, scheduling, brokering • Monitoring the Virtual Organization for failure, intrusion detection, etc.

9. Load Sharing - Overview • Transferring work from a heavily loaded node to a lightly loaded node • Purpose: To improve application performance • Transferring processes not suitable for fine-grain parallelism • Also known as: Load Balancing, Process Migration.

10. Load Sharing Issues • Criteria for heavily-loaded, lightly loaded • Measuring load (policy, implementation) • Exchanging information about load, state • Which jobs to transfer • When to transfer (new processes only, already-running processes)

11. Load Sharing in the Grid • “A Performance Oriented Migration Framework for the Grid”, Vadhiyar and Donngarra • Part of the GrADS project – Grid Application Development System – based at Univ. of Tennessee and other institutions • Designed for long-running computations

12. Load Sharing in the Grid - 2 • Basic idea – the load sharing system can run a performance model of a computation to estimate running time and resource requirements. • Application programmer is responsible for providing performance model for the application, and hooks to stop application, checkpoint state, and re-start application. • Based on MPI Programming Library, Globus Toolkit

13. Load Sharing in the Grid - 3 • Before application begins, Application Manager runs performance model to predict execution times, number of processors. • Determines whether an appropriate set of processors is available, schedules jobs • Monitors process of application as it runs

15. Load Sharing in the Grid - 4 • Load sharing can occur if • Application progress is delayed • Additional resources become available • App Manager sends message to application so it will • Checkpoint • Stop computation • Re-start on new collection of nodes

17. Research Directions • Load sharing on the Grid: • There’s a large body of pre-Grid research of load balancing in distributed systems • Can the results of this research be used to design load balancing systems for the Grid

18. Load Balancing and Grid Computing David Finkel Computer Science Department Worcester Polytechnic Institute