Grid-Computing

Grid computing is a form of distributed computing whereby a supercomputer is created from a cluster of loosely coupled computers. It was designed to solve computationally intensive problems. Today, grid computing is in many areas, some of them commercially, used, for example in pharmaceutical research and economics, electronic commerce and Web services. It is also used for risk management in the structural dynamics and in financial management.

Conventional computer clusters, grid computing differs in the much looser coupling, the heterogeneity and the geographical distribution of the computers. Furthermore, a Grid is usually intended for a specific application and often uses standardized libraries and middleware.

• 2.1 General
• 2.2 OGSA
• 2.3 Virtual Organisations
• 2.4 Implementations / grid middleware
• 2.5 hardware

• 3.1 EGEE / EGI
• 3.2 nordugrid
• 3.3 XtreemOS
• 3.4 neuGRID
• 3.5 National Grid Infrastructure Initiatives

Basics

Definition

The first attempt at a definition comes from Ian Foster and Carl Kesselman in the book " The Grid: Blueprint for a New Computing Infrastructure ":

As this definition was written before the actual emergence of Grids, it was refined by Ian Foster in the second edition of the book clearly once again

The main difference to the original definition is that the sharing of resources is determined by virtual organizations. These play a central role in the current implementations of Grids. Even now are no longer only high-performance computer designated as a resource but general resources, such as physical experiments.

There are other tests of a uniform definition in the literature (cf. ).

Falsely are systems such as cluster computing, peer-to -peer computing, meta computing or distributed computing in particular information (eg, SETI @ home) Grid Systems. Although these have aspects of grid computing, but missing key points in order to be called Grid. This problem Ian Foster has taken in a clear 3- point checklist. The properties of a Grid system are defined in a few words as follows:

Origin

Concepts for the distribution of computationally intensive tasks it has already been given in the 60s. Most of today's research on Grid systems has its origin. Early experiments with high-speed networks In this context, the projects FAFNER and I-WAY be mentioned.

The term Grid has its origin in the comparison of this technology to the electricity grid ( Power Grid ). Thus, the Grid, a user just simply provide resources such as computing power or storage capacity via the Internet, as it is possible to obtain electricity from a wall outlet. The user submits his order via standardized interfaces to the grid, after which the resource allocation is done automatically.

I-WAY

The I -WAY project (Information Wide Area Year ) was carried out in 1995 within the gigabit test bed at the University of Illinois, where 17 bodies were merged in the U.S. and Canada to a high-speed network. It had for its object to connect various supercomputers using existing networks. I-WAY has supported the development of the Globus project to a large extent.

Objective

The motivating objective which has led to the development of the grid technology, was the joint, coordinated use of resources and the joint solution of problems inside of dynamic -border, virtual organizations (see ). This means, according to the definition of accounting rules and rights direct access to, for example, computing power, applications, data or instruments should be jointly made ​​possible. A virtual organization (VO) is in this context is a dynamic association of individuals and / or institutions who share common goals in the use of the grid. Although the focus of many works is with the distributed computing, yet the primary goal is analogous to the emergence of the Internet, to develop a unified, global grids.

Classification

Roughly speaking, can be divided into classes grid computing, such as

• Computational grids ( grid computing ): access to distributed computing resources
• Data grid ( datagrid ): access to distributed databases
• Resource grids
• Service grids
• Knowledge grids

The class of the computing grid is comparable to the power grid, so the grid: this provides the consumers of computing power to connect to the network computing ago, similar to the power consumers to the electricity supply network. There is everything that happens behind the plug, hidden to the consumer; he simply uses the service offered.

In the class of the data grid cooperate not only the (high performance ) computer systems of the parties to provide computing power, but also data sets to be linked. Access to such grids usually provides a Grid portal.

In addition, the provision of network resources " gridifiziert ", that is an automatic selection from a pool of resources is made ​​on the basis of specific QoS parameters. Ideally, the choice of resources should be application- driven, be so dependent on the application in computing grid or data grid.

Architecture and implementations

General

On the Architecture of a Grid, there are several concepts. Any known concept is peculiar that there must be a coordinating body for the agglomeration of computing power and for merging the partial services except the inquirers instance and provided by there actual performance requirement. In addition, a strict hierarchy is required, which allows or excludes the agglomeration of computing power according to objective criteria. Any computer connected to the "grid" is a the other computers first hierarchically equivalent unit (Peer -2-Peer ).

The typical tasks in which grid computing offers itself as a strategy are those that overwhelm the performance of individual computers. These include the integration, analysis and presentation of very large amounts of data from scientific and medical research. In the routine, the techniques are also applied in meteorology and computationally intensive simulations in industry. In particular, the particle physics with large-scale experiments (eg, the Large Hadron Collider) as a scientific application is a pioneer in the development and establishment of Grid technologies.

The typical problems involved in grid computing with it, are the rising costs as part of the available capacity for coordination. Therefore, the processing power increases because of the never coordination expense linearly with the number of the computers involved. This aspect occurs in complex numerical tasks in the background.

OGSA

A possible software architecture for Grids is co-developed by Ian Foster Open Grid Services Architecture ( OGSA ). This is described in approaches already in its predecessor, the Open Grid Services Infrastructure ( OGSI ). Their basic idea is the representation of involved components (computers, storage space, microscopes, ... ) as Grid services in an open component architecture.

With the convergence of Web services of the W3C and the OGSA standards of the Open Grid Forum ( OGF ) Grid services were on a technical level, are as described for example in Globus Toolkit 4 implemented to Web services, which of the technical functionalities Grid middleware provide. OGSA proposes the use of WSRF ( Web Services Resource Framework) as a basic building block for service grid in this context. So get the web services, their use enables uniform access method to the individual services of a grid, additionally a condition: You will be with stateful (English stateful ) ( As files, Java objects or POJOs, records in a database ) Resources associated. This allows only perform functions that span multiple transactions. The combination of web services with such a stateful resource forms a so-called WS- Resource.

• OGF
• OGSA

Virtual organizations

A key and hardware-independent concept behind the Grid philosophy is that of the virtual organizations (VO, see there). Here are resources (or services ) dynamically allocated to virtual organizations.

Implementations / grid middleware

• G - Eclipse
• Globus Toolkit
• Unicore
• GLite
• Sun Grid Engine

Hardware

Practically speaking, you need to hardware nothing more than a computer with a network connection. On this grid computers software takes over the release of a subtask, which was, for example, with the help of software, which can split a big task into a number of subtasks for all nodes in the grid and the partial results summarizing again generated.

Scientific Projects

EGEE / EGI

The Gone in March 2006 to the end of EGEE project ( Enabling Grids for E- sciencE, formerly Enabling Grids for E -science in Europe) is the largest grid project of the European Union, now with worldwide use. Under the name EGEE2 it continues since April 2006. The project was funded in the first phase of the EU with 32 million euros and is the world 's largest Grid infrastructure project in 2010 was replaced by EGEE European Grid Ifrastructure ( EGI).

The participating institutions other CERN ( Switzerland ), Karlsruhe Institute of Technology (KIT, Germany ), Rutherford Appleton Laboratory ( RAL, UK ), Istituto Nazionale di Fisica Nucleare ( INFN, Italy) and Academia Sinica ( ASCC, Taiwan ). See also: Enabling Grids for E- sciencE.

Nordugrid

The open grid nordugrid based on the ARC grid middleware has emerged from a merger of five Scandinavian institutions.

XtreemOS

Building and Promoting a Linux -based Operating System to support Virtual Organizations for Next Generation Grids is a project that is funded under the 6th Framework Programme of the European Union. In addition to 17 European project partners, two are involved from China to XtreemOS. It was launched in July 2006 and will run for four years.

NeuGRID

NeuGRID is a project that was funded by the European Union under the 7th Framework Programme. Its infrastructure enables the scientific research into neurodegenerative diseases.

National Grid Infrastructure Initiatives

As in several other countries (eg U.S.: Cyber ​​Infrastructure, UK: UK e-Science / OMII, NL: BIG- Grid) is also available in Germany and Austria Grid initiatives to develop a national grid infrastructure.

• German Grid Initiative
• Austrian Grid