Grid Computing – Now and future development

There are lots of problems and questions need to be solved in academy. It is time consuming for scientists to solve the problems even they use supercomputer to compute the result. However, the grid computing do a great job on solving some complicated problems that should be done by supercomputer traditionally. Although grid computing is not so popular in normal computer users compare with cloud computing and cluster computing, grid computing is quite popular on solving some professional problems such as scientific and mathematical questions.

What is Grid Computing?

Grid computing refers to the combination of computer resources from heterogeneous computing device to reach a common goal. Although grid computing is a kind of distributive computing, it is different from cluster computing-based systems. In grid computing, each device can be widely spread.  The devices need not be the same computational architecture. They need middleware to divide and allocate the job that the devices are responsible to compute through network which is typically the Internet or Ethernet. There are several big projects using the grid computing to solve the problems, for example, Genome@Home, Folding@Home, MilkyWay@Home etc. All these projects contribute a lot to human society.

Advantage of using Grid Computing

There are many advantages to apply gird computing in scientific, mathematical and academic problems through volunteer computing. This means the people, who share their computing power from their device, will not get any money from institutes. It is voluntary base. The institute does not need to pay a lot to obtain a great computing power. Also, the cost of combining multiple processors in a normal Personal Computer is far lower than cost of making a tailor-made supercomputer due to the mass production of normal processors lower the cost of normal CPU. With grid computing, multiple devices can produce similar computing resource as multi-processor supercomputers. Hence, cost is relatively low compare with setting up a tailor-made super computer to solve the problems. Besides, the institutes need not to have lots of space to distributive computing-based system or a super computer. The maintenance cost can also be saved. 

Disadvantage of using Grid Computing

Although grid computing can lower the cost of computation, there are some restrictions. Firstly, as the devices in grid computing may not have stable and instant connections with other devices, the computation should be dived in the part that is highly parallel and can be solved independently. This increases the difficulty on designing the programme. Moreover, grid computing is mainly relying on voluntary computing. This limits the maximum computing power of grid computing.

Future Development

In past few years, grid computing mainly relies on the CPU as the main processing source. More and more computing power is made by the Graphic Processing Unit (GPU) and the Cell processor in the PlayStation 3 (PS3). These processors have higher Floating Point Operations (FLOPS), which is the main operation of scientific problems, than the traditional CPU in traditional PC. As GPU is not always have heavy work, there are more idling time compare with the CPU. Also, it is impossible for the gamers to play with their PS3 all the day. That means more available computing power. Therefore, the computing power from the GPU and PS3 can be easily obtained compare with CPU. In the future, there are more PCs equipped with high computing power GPU to fulfill the needs of multimedia. With certain promotion of grid computing, it is believed that grid computing will provide a considerable computing power and recycle more idling computing power.

Secondly, more devices have high computing power, such as mobile phones, tablets and Home Theatre PCs (HTPC). Their computing power is increasing while their power consumption is decreasing. They can be the potential devices of gird computing. Plus, development of high speed connection to the Internet through traditional connection and wireless connection makes grid computing easier than before. It is believed that more computing devices can join the grid computing and share the idling computing power. At the same time, the institutes can join their supercomputers and mainframe to speed up the processing. This may form an ultimate virtual supercomputer.

In shorts, grid computing will still develop to help scientists to study difficult problems in future. With more computing devices are available to join grid computing model. The potential of developing grid computing model is unlimited. Also, it is better to promote people to join the grid computing to share their idling computing power in order to greatly boost up computing power of grid computing. Let their devices to join the grid computing is better than idling their devices and do nothing. At least, there are some contributions to society. It is no doubt that developing grid computing is beneficial to us.

Reference

1.    Grid computing – Wikipedia. Retrieved October 2011 from Wikipedia: http://en.wikipedia.org/wiki/Grid_computing

2.    Folding@home. Retrieved October 2011 from Folding@home http://folding.stanford.edu/English/Main

3.    Folding@home – Wikipedia. Retrived October 2011 from http://en.wikipedia.org/wiki/Folding@Home