Resource-Aware Load Balancing of Parallel Applications

Resource-Aware Load Balancing of Parallel Applications

Eric Aubanel (University of New Brunswick, Faculty of Computer Science, Canada)
DOI: 10.4018/978-1-60566-184-1.ch002
OnDemand PDF Download:
$37.50

Abstract

The problem of load balancing parallel applications is particularly challenging on computational grids, since the characteristics of both the application and the platform must be taken into account. This chapter reviews the wide range of solutions that have been proposed. It considers tightly coupled parallel applications that can be described by an undirected graph representing concurrent execution of tasks and communication of tasks, executing on computational grids with static and dynamic network and processor performance. While a rich set of solution techniques have been proposed, there has not been of yet any performance comparisons between them. Such comparisons will require parallel benchmarks and computational grid emulators and simulators.
Chapter Preview
Top

Background

Since heterogeneous computing is a broad research area and there are many definitions of computational grids, we define the characteristics of the applications and platforms that we are considering.

Key Terms in this Chapter

Resource-Aware Load Balancing: Assignment of tasks to a computational platform, taking into account its heterogeneous characteristics, with the objective of balancing the work assigned to processors. The tasks may be independent, or may have interdependencies described in a graph or digraph. In the latter case an additional objective is to minimize communication between processors. The assignment may be static, that is decided before tasks execute, or dynamic, that is tasks are assigned as conditions change.

Dynamic Load Balancing: Assignment of tasks to a computational platform, with the objective of balancing the work assigned to processors. When the tasks are interdependent an additional objective is to minimize communication between processors. The load balance is monitored during execution of the application, and the assignment is modified if required, while trying to minimize the communication cost of migrating tasks.

High Capacity Computational Grid: A grid platform dedicated to high performance computing (HPC) applications and generally consisting of an aggregation of multiple HPC clusters. These grids are space-shared, that is processors only execute a single user’s process(es) at a time. The clusters may be located in one organization, or they may be more spread out geographically, which can pose additional problems of high latencies.

Static Load Balancing: Assignment of tasks to a computational platform, before execution takes place, with the objective of balancing the work assigned to processors. When the tasks are interdependent an additional objective is to minimize communication between processors. This assignment is then fixed for the duration of the computation.

Graph Partitioning: In the context of parallel computing, the partitioning of a graph (representing concurrent execution of tasks and communication between tasks) into disjoint subgraphs, where the communication between subgraphs is minimized and the load is roughly balanced.

High Throughput Computational Grid: A grid platform that aims to take advantage of spare compute cycles on processing nodes. These nodes may be dedicated to multiple users at the same time and hence are time-shared. These grids can consist of local or global networks of computers, and can be either within one organization or be simply a collection of individual computers on the Internet.

Heterogeneous Computing: Design and implementation of algorithms for computing platforms with heterogeneous characteristics, such as processor and network performance.

Complete Chapter List

Search this Book:
Reset
Editorial Advisory Board
Table of Contents
Foreword
Ruth E. Shaw
Preface
Emmanuel Udoh, Frank Zhigang Wang
Acknowledgment
Emmanuel Udoh
Chapter 1
Emmanuel Udoh, Frank Zhigang Wang, Vineet R. Khare
This chapter presents a historical record of the advent of Grid with a recourse to some basic definitions commonly accepted by most researchers. It... Sample PDF
Overview of Grid Computing
$37.50
Chapter 2
Eric Aubanel
The problem of load balancing parallel applications is particularly challenging on computational grids, since the characteristics of both the... Sample PDF
Resource-Aware Load Balancing of Parallel Applications
$37.50
Chapter 3
Enis Afgan, Purushotham Bangalore
Grid computing has emerged as the next generation computing platform. Because of the resource heterogeneity that exists in the grid environment... Sample PDF
Assisting Efficient Job Planning and Scheduling in the Grid
$37.50
Chapter 4
Kuo-Chan Huang, Po-Chi Shih, Yeh-Ching Chung
Most current grid environments are established through collaboration among a group of participating sites which volunteer to provide free computing... Sample PDF
Effective Resource Allocation and Job Scheduling Mechanisms for Load Sharing in a Computational Grid
$37.50
Chapter 5
Tevfik Kosar
As the data requirements of scientific distributed applications increase, the access to remote data becomes the main performance bottleneck for... Sample PDF
Data-Aware Distributed Batch Scheduling
$37.50
Chapter 6
Gianni Pucciani, Flavia Donno, Andrea Domenici, Heinz Stockinger
Data replication is a well-known technique used in distributed systems in order to improve fault tolerance and make data access faster. Several... Sample PDF
Consistency of Replicated Datasets in Grid Computing
$37.50
Chapter 7
Ming Wu, Xian-He Sun
Rapid advancement of communication technology has changed the landscape of computing. New models of computing, such as business-on-demand, Web... Sample PDF
Quality of Service of Grid Computing
$37.50
Chapter 8
QoS in Grid Computing  (pages 75-83)
Zhihui Du, Zhili Cheng, Xiaoying Wang, Chuang Lin
This chapter first summarizes popular terms of QoS related concepts and technologies in grid computing, including SLA, End-to-End QoS Provision and... Sample PDF
QoS in Grid Computing
$37.50
Chapter 9
Kris Bubendorfer, Ben Palmer, Ian Welch
A Grid resource broker is the arbiter for access to a Grid’s computational resources and therefore its performance and functionality has a... Sample PDF
Trust and Privacy in Grid Resource Auctions
$37.50
Chapter 10
Sandro Fiore, Alessandro Negro, Salvatore Vadacca, Massimo Cafaro, Giovanni Aloisio, Roberto Barbera
Grid computing is an emerging and enabling technology allowing organizations to easily share, integrate and manage resources in a distributed... Sample PDF
An Architectural Overview of the GRelC Data Access Service
$37.50
Chapter 11
Man Wang, Zhihui Du, Zhili Cheng
Resource Management System (RMS), which manages the Grid resources and matches the applications’ requests to the proper resources, is one of the... Sample PDF
Adaptive Resource Management in Grid Environment
$37.50
Chapter 12
Vineet R. Khare, Frank Zhigang Wang
The need for a dynamic and scalable expansion of the grid infrastructure and resources and other scalability issues in terms of execution efficiency... Sample PDF
Bio-Inspired Grid Resource Management
$37.50
Chapter 13
Yuhui Deng, Frank Zhigang Wang, Na Helian
Storage Grid is a new model for deploying and managing the heterogeneous, dynamic, large-scale, and geographically distributed storage resources.... Sample PDF
Service Oriented Storage System Grid
$37.50
Chapter 14
Dominic Cherry, Maozhen Li, Man Qi
This chapter presents MediaGrid, a distributed storage system for archiving broadcast media contents. MediaGrid utilizes storage resources donated... Sample PDF
A Distributed Storage System for Archiving Broadcast Media Content
$37.50
Chapter 15
Maozhen Li, Man Qi, Bin Yu
The computational grid is rapidly evolving into a service-oriented computing infrastructure that facilitates resource sharing and large-scale... Sample PDF
Service Discovery with Rough Sets
$37.50
Chapter 16
Irfan Habib, Ashiq Anjum, Richard McClatchey
Due to some barriers to adoption we have not seen a proliferation of Grid Computing technologies throughout e-Science or other domains. This chapter... Sample PDF
On the Pervasive Adoption of Grid Technologies: A Grid Operating System
$37.50
Chapter 17
Kurt Vanmechelen, Jan Broeckhove, Wim Depoorter, Khalid Abdelkader
As grid computing technology moves further up the adoption curve, the issues of dealing with conflicting user requirements formulated by different... Sample PDF
Pricing Computational Resources in Grid Economies
$37.50
Chapter 18
Rosario M. Piro
Large, geographically distributed and heterogeneous computing infrastructures, such as the Grid, often span multiple organizations and... Sample PDF
Resource Usage Accounting in Grid Computing
$37.50
Chapter 19
Frans Arickx, Jan Broeckhove, Peter Hellinckx, David Dewolfs, Kurt Vanmechelen
Quantum structure or scattering calculations often belong to a class of computational problems involving the aggregation of a set of matrices... Sample PDF
Grid-Based Nuclear Physics Applications
$37.50
Chapter 20
Gabriel Aparicio, Fernando Blanco, Ignacio Blanquer, César Bonavides, Juan Luis Chaves, Miguel Embid, Álvaro Hernández
In the last years an increasing demand for Grid Infrastructures has resulted in several international collaborations. This is the case of the EELA... Sample PDF
Developing Biomedical Applications in the Framework of EELA
$37.50
Chapter 21
Gerald Schaefer, Roger Tait
Efficient approaches to computationally intensive image processing tasks are currently highly sought after. In this chapter, the authors show how a... Sample PDF
Distributed Image Processing on a Blackboard System
$37.50
Chapter 22
Daniele Andreotti, Armando Fella, Eleonora Luppi
The BaBar experiment uses data since 1999 in examining the violation of charge and parity (CP) symmetry in the field of high energy physics. This... Sample PDF
Simulated Events Production on the Grid for the BaBar Experiment
$37.50
Chapter 23
Diego Liberati
A framework is proposed that creates, uses, and communicates information, whose organizational dynamics allows performing a distributed cooperative... Sample PDF
A Framework for Semantic Grid in E-Science
$37.50
Chapter 24
Roberto Barbera, Valeria Ardizzone, Leandro Ciuffo
The Grid INFN virtual Laboratory for Dissemination Activities (GILDA) is a fully working Grid test-bed devoted to training and dissemination... Sample PDF
Grid INFN Virtual Laboratory for Dissemination Activities (GILDA)
$37.50
Chapter 25
Dirk Gorissen, Tom Dhaene, Piet Demeester, Jan Broeckhove
The simulation and optimization of complex systems is a very time consuming and computationally intensive task. Therefore, global surrogate modeling... Sample PDF
Grid Enabled Surrogate Modeling
$37.50
Chapter 26
Patrik Skogster
Grid computing is becoming as essential part of different business analysis. In traditional business computing infrastructures data transfer occurs... Sample PDF
GIS Grids and the Business Use of GIS Data
$37.50
Chapter 27
Gokop Goteng, Ashutosh Tiwari, Rajkumar Roy
The emerging grid technology provides a secured platform for multidisciplinary experts in the security intelligence profession to collaborate and... Sample PDF
Grid Computing: Combating Global Terrorism with the World Wide Grid
$37.50
Chapter 28
Salvatore Scifo
This chapter focuses on the efforts to design and develop a standard pure Java API to access the metadata service of the EGEE Grid middleware, and... Sample PDF
Accessing Grid Metadata through a Web Interface
$37.50
Chapter 29
Jyotsna Sharma
Efforts in Grid Computing, both in academia and industry, continue to grow rapidly worldwide for research, scientific and commercial purposes.... Sample PDF
Grid Computing Initiatives in India
$37.50
Chapter 30
Hai Jin, Li Qi, Jie Dai, Yaqin Luo
A grid system is usually composed of thousands of nodes which are broadly distributed in different virtual organizations. Owing to geographical... Sample PDF
Dynamic Maintenance in ChinaGrid Support Platform
$37.50
About the Contributors