Grid Computing: Resource Management Thesis Problem Statement

1

• Grid Computing Resource ManagementThesis
  Problem Statement
• Francisco CastellanosGraduate Student, Computer
  EngineeringFairfield University

2
Contents

• Introduction
• Problem Statement
• Approach
• Motivation
• Work In Progress
• References

3
Introduction

• What is Grid Computing?
• Grid computing is a kind of parallel computing
  that enables the sharing, selection, and
  aggregation of geographically distributed
  autonomous resources, at runtime.1
• One of the services that a Grid can provide is a
  computational service.1

4
Problem Statement

• We propose to address a sub-problem in the area
  of Computational Grids concerned with resource
  management.
• As compute servers become available (idle), they
  run a screen saver.
• The screen saver informs a central Look-Up
  Server (LUS) and volunteers the server to the
  GRID.
• The server then downloads a job, performs some
  computation, and returns the results back to the
  LUS.

5
Problem Statement

• Goals
• Make use of the CS (Compute Server) during its
  idle state.
• Minimize the intrusion into the CS desktop.
• Execute a trusted job and to return the result
  from the CS to the LUS using a secure channel.
• Creation of a portable screen saver that can be
  downloaded on demand.

6
Approach

• Problem Partitions
• Detection of CS Idle-state initiation and
  termination.
• Screensaver deployment.
• Computation task deployment.
• Communication between CS and LUS (Benchmark data
  and job status)
• Leasing

7
Approach

• Proposal
• Use and customization of Saverbeans SDK.
  SaverBeans Screensaver SDK is a Java screensaver
  development kit, enabling developers to create
  cross-platform screensavers.
• Use of Java Web Start for Screensaver deployment
  and Compute Server deployment.
• Create a communication protocol between CS and
  LUS. Information includes benchmark data, load,
  and processing state.
• Implement leasing process.

8
Motivation

• Work stations are idle 14 hours or more per work
  day.
• Waste of CPU cycles, Storage Capacity and network
  bandwidth.
• Success of the SETI project. A major deficiency
  of SETI is that it is not portable and generally
  only available to accelerate a single
  application.
• Grid promises to make it possible for scientific
  collaborations to share resources on an
  unprecedented scale, and for geographically
  distributed groups to work together in ways that
  were previously impossible.2

9
Work in progress

• Screensaver for Linux and Windows.

10
Work in progress

• Screensaver deployment process for Linux and
  Windows with Java Web Start.

11
Work in progress

• Integration of Screensaver and Compute server for
  Linux.

Compute Server
Xscreensaver-demo
12
Work in progress

• Job partitioning Mandelbrot set.

13
References

• 1 Mark Baker, Rajkumar Buyya and Domenico
  Laforenza Grids and Grid technologies for
  wide-area distributed computing,
  http//www.gridbus.org/papers/gridtech.pdf,
  SOFWARE-PRACTICE AND EXPERTICE, 2002, John Wiley
  and Sons,Ltd.
• 2 Ian Foster The Grid A new Infrastructure
  for 21st Century Science, American Institute of
  Physics, 2002, http//www.aip.org/pt/vol-55/iss-2/
  p242.html

14

• END