On incorporating differentiated levels of network service into GridSim

https://doi.org/10.1016/j.future.2006.10.006Get rights and content

Abstract

Grid computing technologies are increasingly being used to aggregate computing resources that are geographically distributed. Commercial networks are being used to connect these resources, and thus serve as a fundamental component of Grid computing. Since these Grid resources are connected over a shared infrastructure, it is essential that we consider the effects of using this shared infrastructure during simulations. In this paper, we discuss how new additions to the GridSim simulation toolkit can be used to explore network effects in Grid computing. We also investigate techniques to incorporate differentiated levels of service, background traffic and the collection of information from the network during runtime in GridSim. As a result, these features enable GridSim to realistically model Grid computing experiments.

Introduction

Grid computing has emerged as the next-generation parallel and distributed computing methodology, which aggregates dispersed heterogeneous resources for solving various kinds of large-scale parallel applications in science, engineering and commerce [10]. In order to evaluate the performance of a Grid environment, we need to conduct repeatable and controlled experiments, which are difficult due to the Grid’s inherent heterogeneity and its dynamic nature. Additionally, Grid testbeds are limited, and creating an adequately-sized testbed is expensive and time consuming. Moreover, it requires the handling of different administration policies at each resource. Due to these reasons, it is easier to use simulation as a means of studying complex scenarios.

The GridSim toolkit [5] has been developed to overcome the above problems. It is a Java-based discrete-event Grid simulation package that provides features for application composition, information services for resource discovery, and interfaces for assigning applications to resources. GridSim also has the ability to model the heterogeneous computational resources of various configurations. The GridSim toolkit has been applied successfully to simulate a Nimrod-G-like [6] Grid resource broker and to evaluate the performance of deadline and budget constrained cost- and time-optimization scheduling algorithms.

Communication networks serve as fundamental components of Grid computing. Resources, connected over commercial networks, share bandwidth with other users. A realistic simulation of Grid environments should include the effects of sending data over shared communication lines. Earlier versions of GridSim did not have the ability to specify a network topology, nor the functionality to connect resources through network links during the experiment. Resources and Grid users had all-to-all connections with specifiable bandwidths. Hence, the simulations did not capture the entire details of an actual Grid testbed.

In this work, GridSim has been extended to address the above problems by enhancing the ability to simulate realistic network models by: (1) allowing users to create a network topology, (2) packetizing data into smaller chunks for sending over a network, (3) generating background traffic, and (4) incorporating different levels of service for sending packets.

The rest of this paper is organized as follows: Section 2 provides some relevant background on GridSim. Section 3 presents the design and implementation of the network additions to GridSim, while Section 4 illustrates the use of GridSim for simulating a Grid computing environment. Section 5 mentions related work. Finally, Section 6 concludes the paper and suggests some further work to be done on GridSim network models.

Section snippets

Background

There has been significant work done in the past to incorporate more functionality and extensibility into GridSim ver3.0, such as extending the GridSim infrastructure to support advance reservation as discussed in [26]. This allows resources to have their own schedulers and policies in reservation-based systems. However, no work has been done on improving the existing network model. Therefore, in the newer GridSim release, a new package is incorporated to enhance the capabilities of the

Design and implementation of the GridSim network

The flow of information among GridSim entities happens via their Input and Output (I/O) entities. Upon creating an entity with a specified bandwidth, GridSim creates a new instance of the Input and Output classes, and links them to the new entity. Hence, data sent by an entity goes through its Output entity, and is received by other entities via their Input entities.

The use of separate entities for I/O provides a simple mechanism for GridSim entities to communicate with each other, and allows

Experiment Aim

The main aim of this experiment is to show the behavior of the network components and the packet scheduling algorithms implemented in GridSim. Hence, we are trying to look at:

  • how background traffic can affect network loads and overall packet execution time; and

  • how differentiated QoS levels for packets can help in a heavy load situation.

In order to conduct this experiment, we use a network topology based on the EU DataGrid TestBed I [28]. The topology from this production Grid is chosen because

Related work

Simulation is widely used in the networking research area. Examples of such simulators include NS-2 [29], DaSSF [17], OMNET++ [30] and J-Sim [14]. Though their support for network protocols is extensive, they are not targeted at studying Grid computing. This is because simulating Grids requires modeling the effects of job scheduling algorithms on Grid resources, and investigating users’ QoS requirements for application processes. In addition, we believe that simulating TCP and UDP connections

Conclusion and further work

The network serves as a fundamental component in Grid computing, since resources and users are connected over a network topology with shared bandwidth. Previously, GridSim did not have the ability to specify a network topology or the functionality to connect resources through network links during experiments. In this work, modifications to an existing network architecture have been incorporated into GridSim to address the above problems.

With the addition of this network functionality, users can

Software availability

The latest version of the GridSim toolkit with source code and examples, can be downloaded from the following website: http://www.gridbus.org/gridsim/.

Acknowledgements

We thank Uros Cibej for his work on implementing the functionality of creating a network topology from a file. We also thank CS Yeo, Hussein Gibbins and anonymous reviewers for their comments.

Anthony Sulistio is a Ph.D. student at the Department of Computer Science and Software Engineering (CSSE), University of Melbourne, Australia. He received his B.E. and M.S.S.E. degree from University of Melbourne in 2001 and 2002 respectively. His main research interests are in grid computing, grid simulation, parallel programming and software engineering.

References (33)

  • R. Wolski et al.

    The network weather service: A distributed resource performance forecasting service for metacomputing

    The Journal of Future Generation Computing Systems

    (1999)
  • K. Aida et al.

    Performance evaluation model for scheduling in a global computing system

    The International Journal of High Performance Computing Applications

    (2000)
  • W.H. Bell et al.

    Optorsim—a Grid simulator for studying dynamic data replication strategies

    The International Journal of High Performance Computing Applications

    (2003)
  • S. Blake, D. Black, M. Carlson, E. Davies, Z. Wang, W. Weiss, RFC 2475: An architecture for differentiated service,...
  • Bricks: A performance evaluation system for Grid computing scheduling algorithms. Available:...
  • R. Buyya et al.

    Gridsim: A toolkit for the modeling and simulation of distributed management and scheduling for Grid computing

    The Journal of Concurrency and Computation: Practice and Experience

    (2002)
  • R. Buyya, D. Abramson, J. Giddy, Nimrod-G: An architecture for a resource management and scheduling system in a global...
  • H. Casanova, SimGrid: A toolkit for the simulation of application scheduling, in: Proc. of the First IEEE/ACM...
  • A.J. Demers, S. Keshav, S. Shenker, Analysis and simulation of a fair queueing algorithm, in: Proc. of the ACM...
  • E. Elmroth, P. Gardfjall, Design and evaluation of a decentralized system for Grid-wide fairshare scheduling, in: Proc....
  • I. Foster et al.

    Globus: A metacomputing infrastructure toolkit

    The International Journal of Supercomputer Applications and High Performance Computing

    (1997)
  • I. Foster

    Globus toolkit version 4: Software for service-oriented systems

  • S.J. Golestani, A self-clocked fair queueing scheme for broadband applications, in: Proc. of IEEE INFOCOM’94, Toronto,...
  • J-Sim. Available: http://www.j-sim.org/...
  • A. Legrand, L. Marchal, H. Casanova, Scheduling distributed applications: The simGrid simulation framework, in: Proc....
  • Cited by (76)

    • Rules discovery in fuzzy classifier systems with PSO for scheduling in grid computational infrastructures

      2015, Applied Soft Computing Journal
      Citation Excerpt :

      In this work, the proposed learning strategy for FCSs, KARP, will be studied to improve the performance of FCS metaschedulers in grid computing. The fuzzy FCS metascheduler with KARP learning is evaluated through simulations with Alea software [31], a simulation toolkit based on GridSim [32] that lets the use of grid scenarios and traces from existing facilities for scheduling strategies evaluation. Specifically, the grid scenario is based on a Czech National Grid Infrastructure Metacentrum project [33], whose aim is to support the development of a distributed high-performance computing infrastructure through the joint cooperation of resources from academic and research institutions worldwide.

    • Meta-schedulers for grid computing based on multi-objective swarm algorithms

      2013, Applied Soft Computing Journal
      Citation Excerpt :

      The first topology is a combination between two testbeds. The topology information is based on the EU DataGrid testbed [24] and it is completed with grid resource characteristics from the WWG testbed [5]. The combination of these testbeds is shown in Fig. 4 and in Table 1.

    View all citing articles on Scopus

    Anthony Sulistio is a Ph.D. student at the Department of Computer Science and Software Engineering (CSSE), University of Melbourne, Australia. He received his B.E. and M.S.S.E. degree from University of Melbourne in 2001 and 2002 respectively. His main research interests are in grid computing, grid simulation, parallel programming and software engineering.

    Gokul Poduval is a Post-Graduate student at the Department of Electrical and Computer Engineering (ECE) of the National University of Singapore (NUS). His research interests are in coordinated quality of service (QoS) management in computational and network systems. He obtained his Bachelor’s degree in 2003 at the same university after receiving scholarship from Singapore Airlines and Neptune Orient Lines.

    Rajkumar Buyya is a Senior Lecturer and the Director of the Grid Computing and Distributed Systems Laboratory within the Department of CSSE, University of Melbourne. He has authored/co-authored over 130 papers and technical documents that include three books—Microprocessor x86 Programming, Mastering C++, and Design of PARAS Microkernel. He received B.E, M.E, and Ph.D. degrees from Mysore, Bangalore, and Monash Universities respectively. He was awarded Dharma Ratnakara Memorial Trust Gold Medal for academic excellence in Mysore University. He is currently serving as the Chair of the IEEE Technical Committee on Scalable Computing (TCSC) and Associate Editor of the Journal of Future Generation Computing Systems (FGCS), Elsevier Press, Holland.

    Chen-Khong Tham is an Associate Professor at the Department of Electrical and Computer Engineering (DECE) of the National University of Singapore (NUS). His research interests are in coordinated quality of service (QoS) management in wired and wireless computer networks and distributed systems, and distributed algorithms for resource allocation, machine learning and decision making. He obtained his Ph.D. and M.A. degrees in Electrical and Information Sciences Engineering from the University of Cambridge, United Kingdom, and held a 2004/05 Edward Clarence Dyason Universitas21 Fellowship at the University of Melbourne, Australia.

    View full text