IDC predicts that the worldwide grid computing market will become US$12 billion by 2007.
Also, the grid computing has been more applied in the business world than ever.
According to IBM, the following industries have shown larger potentials for the grid computing solutions so far.
• Financial services: for risk management and compliance
• Automotive: for the acceleration of product development
• Petroleum: for the discovery of oils
Grid computing make a group of computers connected to the network work as one large computer system.
It is traditionally said that mainframes are idle for 40%,
that Unix servers are idol for 90%,
and that PC servers are idol for 95% of time.
Moreover, even their peak hours, 0-15% time of mainframes, 30-50% time of Unix servers, and 70% time of PC servers are unused.
The basic concept of the grid computing is to utilize these wasted resources to virtually create a higher level of computing as shown above.
Although the grid computing primarily distributes processing power to a group of computers, it is also able to distribute data, application, and whatever resources used in the computing system by using the web services standards. Thus, the grid computing will probably evolve in the following way.
So far leading vendors such as IBM and academics predict that grid computing evolve through the following four steps:
• Processing grid
• Data grid
• Service grid
• Virtualization & utility computing
Processing grid mainly collect and utilize the unused CPU power of computers connected to the network. Data grid distributes huge amount of data file over to computers while service grid allows applications to be shared and collaborated with each other. Eventually, grid computing evolves to the utility computing
when it become completely standardized and real-time, and have a pricing system.
