Scheduling in grid environments

As the demand for more powerful computing resources continually increases, especially in areas of science, engineering, and commerce, a myriad of high-performance computing systems, such as supercomputers and computer clusters, have been built with various different architectures. In general, the use of these specialist computing systems is confined to specific groups of people. Moreover, each of these systems is generally restricted to independent use; that is, it is highly unlikely that a user of one system can access other organizations’ systems. A solution to this is grid computing. A grid enables a virtual computing system interconnecting these geographically distributed heterogeneous computing systems with a variety of resources to be constituted. Here, resources refer not only to physical computers, networks, and storage systems but also to much broader entities, such as databases, data transfer, and simulation. The grid creates the illusion that its users are accessing a single, very powerful, and reliable computing system.