Grid computing (grid – lattice network) is a form of distributed computing in which a virtual supercomputer is presented in the form of clusters connected by a network of loosely coupled, heterogeneous computers working together to perform a huge number of tasks.
This technology is used for scientific and mathematical problems requiring significant computing resources. Grid computing is also used in commercial infrastructure solutions for time-consuming tasks such as economic forecasting, development and study of the properties of new drugs.
Grid from the point of view of the network organization is a consistent, open and standards-based environment that provides a flexible, secure, coordinated division of computing and storage resources information that are part of that environment, within a single virtual organization.
The concept of Grid
Grid is a geographically distributed infrastructure, uniting many different types of resources (processors, and long-term memory, storage, and network database), access to which user can get from anywhere, regardless of their location.
The idea of grid computing has arose with the proliferation of personal computers, the development of Internet technologies and packet data based on optical fiber (SONET, SDH and ATM), as well as local area network technologies (Gigabit Ethernet).
Bandwidth communication tools have become sufficient to attract resources, if necessary, another computer. Given that many are connected to the global network of computers spend most of their time idle, and have the resources to solve their daily problems, there is an opportunity to apply their unused resources elsewhere.
Comparison of grid systems and traditional supercomputers
Distributed or grid computing in general entails parallel computing, which is based on conventional computers (with standard processors , storage devices, power supplies, etc.), connected to the network (LAN or WAN) using conventional protocols such as Ethernet.
While conventional supercomputer comprises a plurality of processors connected to the local high-speed bus.
The main advantage of distributed computing is that a single cell of a computer system can be purchased as an ordinary non-specialized computer.
Thus it is possible to obtain substantially the same computational power as in conventional supercomputers, but with much lower cost.