The following pages will provide you with introductory texts about basic and essential themes around grid computing like e.g. concepts, architecture etc. Further, you will find information about several grid middleware as well as examples of use from several D-Grid projects.
Computing of the next generation
A principal reason for the linking-up of computing resources was the user’s demand for distributed work environments. There were lots of arguments for it:
- A user needs computing resources, which are not available to him on the spot, or data lying on external storage systems, or applications, which are not installed on the local desktop computer or calculation / storage capacity beyond his or her own possibilities.
- A group of users shares computing resources for a common purpose.
- Providing central computing resources for many users aims at increasing or allowing for efficiency.
- A provider has computing resources at his disposal which he may make available to other users.
In closed user groups or within a closed organization such distributed working environments have been common standard for many years. The precondition, however, is a well-regulated user and rights management within these environments.
The future of distributed computer applications
Grid computing deals with the organization of distributed computing workflows in a generalized way. Like in the world wide web, where data and information is provided, linked together and exchanged, grid computing as a further development of the world wide web aims at providing and linking any computing services on linked-up but independent computer systems. The start of a program, the use of a specific hardware or data storage runs over the network using a standardized technique. The user does not see a network of distributed computer systems anymore, but is involved in a virtual system which provides all necessary services. To access and to link-up all provided services, there are open technical specifications which are implemented in the form of grid middleware. Usually, the services are offered and administered by independent operators. Ideally, access to these services is as easy for users as access to electrical power via the power grid.
A trend-setting future technology
Meanwhile, the idea of grid computing has been funded nationally and internationally as a trend-setting future technology for about 10 years. In Germany, the building of a grid computing infrastructure has been fostered by the Federal Ministry of Education and Research (BMBF) in cooperation with Science and Economy in the German Grid Initiative (D-Grid) since 2005. Under the patronage of D-Grid, at present (November 2008) 20 projects are involved in building up and using grid technologies. The range of applications spans from physics and engineering sciences to financial services and humanities. In the course of the current D-Grid 3 initiative, cooperative projects will be fostered by economy and science to support the development of novel, professional, sustainable and user-friendly grid services. Reliability and functionality of grid services will be improved and interdisciplinary basic services in the field of knowledge technologies will be developed.
The future has already begun
It is the status quo, that there are first technical solutions for operating grid infrastructures. The use of these technical solutions is distinct, compared to international projects. Most users up to now are scientists, being reliant on the new possibilities of grid computing. Grid users ally to virtual organizations (VO), which are established around specific topics. For each virtual organization, there are contact persons who also administrate the access to the specific organization. Authorization and authentication for access to a virtual organization is carried out by user certificates. Each virtual organization is assigned to computing resources, which can be accessed by the members of the organization.
The question is not whether, but when grid computing will be established
Today, there is no discussion whether grid computing will become accepted or not. The question is: When will it be established? That is a legitimized question because building up a common grid infrastructure is related to very complex problems: technical protocols for grid users need to be set up and implemented, security problems need to be solved, procedures of authorization and authentication need to be introduced, procedures of accounting and payment need to be developed, legal questions need to be solved and new business models are required. The solution of these problems will be related to changes comparable to the introduction of the world wide web almost 20 years ago.