Basics of Grid-Computing

Submitted by grid on Fri, 25/01/2008 - 14:35.

1. De?nition

Grid is an infrastructure, which allows integrated, collaborative use of geographically separated, autonomous resources. Attempts to find a single definition in the literature, however, are very often(see [12, 13, 14]). This problem Ian Foster, a co-founder of the driving of the Grid technology at the University of Chicago, tried to solve and he has compiled a clear 3-point-checklist. The characteristics of a grid system are briefly defined as follow: A Grid…
  • … coordinates resources, which aren’t subordinated to a central authority ...
    A grid coordinates and integrates resources and users of different administrative domains inside the same company or in different countries. In this context the areas of security, billing and membership are reflected (see [16, 17]). Resources can be clusters, mass storage, databases, applications or measuring devices, among others.
  • ... and uses open, standard protocols and interfaces,
    A Grid uses open and general protocols and interfaces that provide basic functions for the authentication, the authorization, resource identification and the access to resources.
  • ... to provide not trivial qualities of services.
    A Grid uses existing resources in a coordinated manner to provide various qualities of services depending, for example, on the response time, throughput, availability or security. Or there will be a remanagement of multiple resource types in order to answer to complex user expectations in such a way that the using of the combined system is significant bigger than the sum of its parts.
This de?nition is useful for the classification of various technologies into a Grid. Very often systems in the areas of “Distributed computing”, “Cluster computing", "peer-to-peer computing" or "meta computing" were misleadingly called Grid Systems, although many aspects of the aforementioned technologies are used in a Grid.

2. Origin

Concepts for distribution of computational intensive tasks have been arised already in the 60s. Most of the current research work on grid systems has its origin in early experiments of high-speed networks such as in the Gigabit - the test bed at the University of Illinois in 1992 [18]. In particular, however, the I-Way-Experiment [19] in the year 1995, in which 17 institutions in the United States and Canada were joined together with a high-speed network, made the Grid Technology Research publicly visible outside the field of network technology. The term "grid" has his origin in the comparison of this technology to the electricity grid ("Power Grid"). According to this, the grid should make resources like e.g. computing power or storage space via internet available as simple as obtaining electricity from a power socket.

FeaturePower GridComputational Grid
Infrastructure Interconnecting different power station and transformer stations over electric cable. Interconnecting different resources over the internet using open standards.
Transparency Origin of electricity to end-users insignificant. Origin of the resource (e.g. computing power) insignificant.
Availability A very good infrastructure of the power grid is by linking almost all households. The use of different platforms and the Internet gives a very good infrastructure.
Accounting The accountinging is based on the amount of electricity used. The accounting is based on the amount of resources used.
Tabelle 1.1.: Comparison: Power Grid and Computational Grid. Based on [20].

3. Objective

The motivating aim that led to the development of grid technology was the common, coordinated usage of resources and the joint solution of problems inside of dynamic virtual organizations (compare [13]). That means that after the definition of accounting rules and rights there should be a direct and collective access enabled to e.g. computing services, applications, data and scientific instruments. A virtual organization (VO) in this context is a dynamic association of individuals and / or institutions, which follows common objectives by using grids. Indeed, the focus of many works lies in the area of distributed computing, though, in analogy to the emergence of the internet, the highest aim is the development of a unified global grid, that could find use in many areas like science, medicine and teaching. (see [21, 20]).

4. Architecture

4.1. Introduction

The result of about a decade of intensive research and development of Grid- Technologies was among others a consensus about the requirements and the architecture of a grid. Also open standard protocols for message communication and controlling were designed to make the basis for further interoperable development. The relevance of such a Basis becomes apparent in picture 2.1.

 Conceptual structure of a grid
Picture 2.1: Conceptual structure of a grid [21]

Protocols and interfaces can be categorized depending on their function. Because of this, architecture of a grid gets often divided functional into various layers. Components on one layer have similar characteristics and can base on features and behaviour of the lower layers.

Fabric. The lowest layer job is used to make a common interface on all possible kinds of resources available. Access by higher layers is granted via standardized processes. All resources on which such a standardized interface is applicable, can be integrated in the grid concept. This contains computers, storage systems, networks or sensors.

Resource and connectivity protocols: The connectivity layer defines the basic communication- and authentication protocols which are needed by the grid. While the communication protocols allow the exchange of files between different resources connected by the first layer, the authentication protocols allow to communicate confidentially and to ensure the identity of the two partners. To this belongs also delegation of rights and methods for unique authentication (single sign-on). On the resources layer, the common access on individual resources is organized. This contains initiation, observation, control, clearance and negotiation of security parameters. Also processor resources get assigned, reserved, observed and controlled. The OGSA [22] is a standing architecture still in development that will take the implementation of this layer in many grid projects. The Globus Toolkit 4 (GT4) presents a popular implementation of OGSA specification and offers software jobs and libraries to realize a grid according to OGSA specification (see downwards).

Collective services: The purpose of this layer is the coordination of multiple resources. Access to these resources doesn’t happen directly but merely via the underlying protocols and interfaces. The jobs of this layer contain among others the creation of a directory service, they supply monitoring, diagnostic and file replication services. Furthermore grid-capable development systems are provided to be able to use popular programming models also in a grid environment.

User applications: To this layer belong all those applications which are operating in the environment of a virtual organization. Jobs of the lower layers get called by applications and can use resources transparently.

4.2. Globus Toolkit 4

The Globus Toolkit of the Globus-Alliance is a Middleware for Grid Systems. Although it is still in development, almost all major grid projects project are based on this toolkit, so even Instant-Grid does. It is therefore shortly described as a reference implementation of the OGSA specification. The Globus project arose from a collaboration between the University of Chicago and the University of Southern California, with the participation of IBM and NASA. It is based among others on the experience of other projects related to Grid-Technology such as Condor[25], Codine / Sun Grid Engine[26], Legion[27], Nimrod[28] and Unicore[29]. The GT4 offers all necessary components for the implementation of Grid systems. This includes areas of security and data-, resources-, and administrative tasks. In addition, it provides interfaces and libraries for popular programming environments. (see picture 4.1).

components of the Globus Toolkit 4
Picture 4.1: Componenten of the Globus Toolkit 4 [30]

Security. This area includes measures for authentication and authorization of users and resources. The communication between the parties is secured and functions for the management of user and group rights are provided.

Data Management. Modules in this section enable distributed data in the grid to locate (RLS), move (GridFTP, RTP, DRS) and to manage (OGSA-DAI).

Execution Management. The tools for execution management deal with the initialization, monitoring, planning and coordination of distributed tasks.

Information Services. Information Service (MDS) and the related components are used for surveillance and the identification of resources and services in the Grid.

Common Runtime. The general run-time environment is used for platform implementation of the above-mentioned services. There are libraries provided by the creation of abstraction layers and simplify the implementation of features based on Web services (WSRF) comfortably. The installation and configuration of GT4 requires because of its complexity, a high degree of expertise and a significant time. A detailed description of the process can be found amongst others, in [31, 32]. One simple way to demonstrate the tool kit provideses the Instant-Grid project.

5. Literatur

  • [12] I. Foster and C. Kesselman, "The grid: Blueprint for a new computing infrastructure," Morgan Kaufmann, Tech. Rep., 1999.
  • [13] I. Foster, C. Kesselman, and S. Tuecke, "The anatomy of the Grid: Enabling scalable virtual organizations," Lecture Notes in Computer Science, vol. 2150, pp. 1–??, 2001, fetched on the 21st of September 2006. [Online]. Available: http: //citeseer.ist.psu.edu/foster01anatomy.html
  • [14] I. Foster, C. Kesselman, J. M. Nick, and S. Tuecke, "The physiology of the grid," IBM Corporation, Poughkeepie, NY 12601, Tech. Rep., June 2002.
  • [15] I. Foster, "What is the grid? a three point checklist," July 2002.
  • [16] I. Foster, C. Kesselman, G. Tsudik, and S. Tuecke, "A security architecture for computational grids," in ACM Conference on Computer and Communications Security, 1998, pp. 83–92, fetched on the 13rd of August 2006. [Online]. Available: http://citeseer.ist.psu.edu/foster98security.html
  • [17] M. Humphrey and M. Thompson, "Security implications of typical grid computing usage scenarios," in Proceedings of Intl Symposium on High Performance Distributed Computing. (HPDC), San Francisco, CA., August 2001, pp. 7–9.
  • [18] C. Catlett, "In search of gigabit applications," Communications Magazine, IEEE, vol. 30, no. 4, pp. 42–51, 1992.
  • [19] T. DeFanti, I. Foster, M. Papka, R. Stevens, and T. Kuhfuss, "Overview of the i-way: Widearea visual supercomputing," in International Journal of Supercomputer Applications, vol. 10, 1996, pp. 123–130.
  • [20] CERN - European Organization for Nuclear Research, "Gridcafé - a brief explanation of the grid," fetched on the 21st of September 2006. [Online]. Available: http://gridcafe.web.cern.ch/gridcafe
  • [21] I. Foster, "The Grid: A New Infrastructure for 21st Century Science," Physics Today, vol. 55, no. 2, pp. 42–47, 2002.
  • [22] P. Janson, J. Dayka, A. Nadalin, N. Nagaratnam, F. Siebenlist, V. Welch, I. Foster, and S. Tuecke, "The security architecture for open grid services," IBM Corporation, Zurich Research Lab, Switzerland and Mathematics and Computer Science Division, Argonne National Laboratory, Argonne, Tech. Rep., July 2002.
  • [23] I. Foster and C. Kesselman, "Globus: A metacomputing infrastructure toolkit," The International Journal of Supercomputer Applications and High Performance Computing, vol. 11, no. 2, pp. 115–128, Summer 1997, fetched on the 21st of September 2006. [Online]. Available: http://citeseer.ist.psu.edu/article/foster96globu.html
  • [24] I. Foster, "A globus primer," An Early and Incomplete Draft, August 2005.
  • [25] M. Litzkow, M. Livny, and M. Mutka, "Condor-a hunter of idle workstations," Distributed Computing Systems, 1988., 8th International Conference on, pp. 104–111, 1988.
  • [26] W. Gentzsch et al., "Sun Grid Engine: Towards Creating a Compute Power Grid," Proceedings of the First IEEE/ACM International Symposium on Cluster Computing and the Grid, 2002.
  • [27] A. Natrajan, A. Nguyen-Tuong, M. Humphrey, M. Herrick, B. Clarke, and A. Grimshaw, "The Legion Grid Portal," Concurrency and Computation: Practice and Experience, vol. 14, no. 13-15, pp. 1365–1394, 2002.
  • [28] R. Buyya, D. Abramson, and J. Giddy, "Nimrod/g: An architecture for a resource management and scheduling system in a global computational grid," in Proceedings of the HPC ASIA2000, the 4th International Conference on High Performance Computing in Asia-Paci?c Region, Beijing, China. IEEE Computer Society Press, USA, 2000.
  • [29] D. Erwin et al., "UNICORE: A Grid computing environment," Concurrency and Com- putation: Practice and Experience, vol. 14, no. 13-15, pp. 1395–1410, 2002.
  • [30] Globus Documentation Project, "The Globus Toolkit 4 Programmer’s Tutorial," fetched on the 21st of September 2006. [Online]. Available: http://gdp.globus.org/ gt4-tutorial
  • [31] ——, "GT4 Admin Guide," fetched on the 22nd of September 2006. [Online]. Available: http://www.globus.org/toolkit/docs/4.0/admin/docbook
  • [32] Leibniz-Rechenzentrum der Wissenschaften, "Installationsanleitung für Globus Toolkit 4.0.1," fetched on the 22nd of September 2006. [Online]. Available: http://www.grid.lrz.de/de/mware/globus/GT4InstallGuide1-0.html