SE 293 (3:0)

Class

GC2005.JPG (708675 bytes)

Instructor

Sathish Vadhiyar

Books

Syllabus

Orientation
  1. Introduction
    1. Motivation
    2. Definitions of Grid Computing
    3. Evolution of the Grid
    4. Differences with similar efforts (Meta, cluster, heterogeneous, Internet)
    5. Examples of usage
    6. Research possibilities / scope in Grid Computing
    7. Thrust area in this course
  2. The Earliest Grid Motivations
    1. High Performance computing across installation sites - the PACX-MPI example
    2. High Throughput computing using non-dedicated workstations - Condor Part 1, Part 2, Condor Publication on Checkpointing Technique
  3. The Building Blocks of Grid
    1. The Globus toolkit
    2. Security - Kherberos vs Globus GSI
    3. Information Services - NWS
    4. Projects over Globus  - e.g. Condor-G
  4. HPC and Grids
    1. Scheduling HPC applications in Grids
      1. AppLeS
      2. Scheduling Parameter sweep applications
      3. Metascheduling
    2. Grid RPC mechanisms Part 1, Part 2
    3. Rescheduling
  5. Advanced Topics
    1. Data Management in Grids
    2. Grid simulation - MicroGrid
    3. Grid Applications
    4. Grid economy
    5. Grid standards and forums - OGSA, GGF and Other topics

Assignments

  1. Simple Metacomputing Middleware
  2. Condor assignment
  3. AppLeS with Jacobi 2D or Fault-tolerant malleable parallel application
  4. Grid FTP Usage

Class Paper Presentations

Students are required to choose an area that is outside their project area topic.

Name Date Presentation Paper(s)
Mrugesh Gajjar October 19 Memory Conscious Task Partition and Scheduling in Grid Environments. Ming Wu, Xian-He-Sun. Grid 2004.
Yadnyesh Joshi October 24 Optimal File-Bundle Caching Algorithms for Data-Grids. Ekow Otoo, Doron Rotem, Alexandru Romosan. SC 2004.
Prashant Khodade October 26 Improved Message Logging versus Improved Coordinated Checkpointing for Fault-tolerant MPI. Pierre Lemarinieret al. Cluster 2004.
Sumanta Mukherjee November 7 Adaptive Parallel Job Scheduling with Flexible CoScheduling. Frachtenberg et. al. ITPDS, vol. 16, no. 11, Nov. 2005.
Saumya Prakash Panda November 9 Mapping Subtasks with Multiple Versions on an Adhoc Grid. Shivle et. al. ISPDC/HeteroPar'04. 2004.
H.A. Sanjay November 14 Chameleon: A Resource Scheduler in a Data Grid Environment. Park et. al. CCGRID 2003.
Major Chandra Sekhar November 16 A Grid Information Service Based on Peer-to-Peer. Puppin et. al. EuroPar-05.
Major Sandip Singh November 21 Hybrid Preemptive Scheduling of MPI applications. Bouteiller et. al. Grid 2004.

Suggested areas for the Presentations:

Final Project

Broad Topics

  1. Checkpointing, Fault tolerance and Migration
  2. Application-specific middleware - Parallel Applications and Algorithm that need to be modified for execution on Grid environments.

Papers to Look at:

  1. Cluster Computing and Grid (CCGrid)
  2. Euro-Par
  3. Grid
  4. High Performance Distributed Computing (HPDC)
  5. ICDCS
  6. IPDPS
  7. Supercomputing (SC)
  8. IEEE Transactions on Parallel and Distributed Systems (ITPDS)
  9. Journal of Parallel and Distributed Computing (JPDC)
  10. Journal of Grid Computing

Projects:

Team Title
Mrugesh Gajjar / Maj. Chandra Sekhar Checkpointing / Fault-tolerance
H.A. Sanjay Performance Prediction
Prashant Khodade / Maj. Sandeep Singh Docking Application on Grid
Saumya Parakash Panda / Sumanta Mukherjee Genetic Algorithm based Scheduling
Yadnyesh Joshi Hybrid Models for Froecasting