Grid Computing is the process of distributing tasks over a wide range of computers. These tasks can vary from very small data storage to highly complicated calculations. Grid computing is not restricted to one particular locality but can be extended over vast geographical areas. Grid computing can also be referred to as the integration of computer resources to accomplish a common goal.
Grid can be imagined to be a distributed environment with heterogeneous systems that consist of enormous data files. The dimensions of a grid are not always the same. Moreover, one single grid can be used for several purposes. Grid computing can be easily distinguished from other traditional systems such as cluster computing in such a way that a grid is more loosely coupled, is heterogeneous and widespread.
Grid computing can also be referred as ‘Distributed Computing’. It is a Parallel Computing technique that depends on computer systems linked to a network through traditional network interfaces. This is in contradiction to a Supercomputer in which a number of processors are connected by a local high speed computer bus.
CPU Scavenging leads to the creation of a grid by taking the dormant resources present in a computer network. CPU Scavenging uses the instruction cycles of a desktop computer which would have been wasted by just waiting for an input from the user or from an input device.
Advantages of Grid Computing
- Idle resources can be efficiently handled by the use of Grid Computing.
- Large, complex problems can be solved in a short span of time.
- Easier to team up with other organizations.
- Grid environments are modular and they do not fail due do the error that occurs at one particular point. Jobs can be restarted automatically when there is a failure.
- Jobs can be done in parallel thus resulting in faster execution.
Disadvantages of Grid Computing
- Need a fast interconnection between the computer resources.
- A few applications may need to be altered to take full advantage of the new model.
- Many groups do not wish to share resources even if it benefits everybody.
- Jobs submitted may be non-interactive thus leading to chaos.
- Grid computing standards and grid software are still in developing state.