How does Cluster Computing Works?

Asked By 0 points N/A Posted on -
qa-featured

I just want to know how does cluster computing works. Is it the same as when you submitted your work,, cluster computing will rip it by pieces and then distribute those pieces to various workstations that disperse through cluster? Or does cluster computing used only one workstation and run the entire task merely as the front work? I hope you can send me some ideas. Thanks.

SHARE
Answered By 0 points N/A #113087

How does Cluster Computing Works?

qa-featured

 

Computer clusters are generally groups of CPUs that work jointly as one unit. These clusters are usually formed so as to run large programs, or help to augment performance.

In practice, this basically signifies that each processor of each computer executes a share of a given duty. Frequently, clusters are used chiefly for computational purposes, rather than taking care of IO-oriented tasks like web service or databases. 

Computer clusters are normally preferred by business and enterprise to give low-cost, but also highly efficient processing capabilities and they are in general considered to be cheaper as compared to a new computer with similar capacity.

Answered By 574215 points N/A #311199

How does Cluster Computing Works?

qa-featured

The term “cluster” refers to independent computers combined into a unified system by means of networking and software. At the most basic level, when two or more computers are used together to solve a problem, it is already considered a cluster. Cluster computing is used for high performance and high availability computing. It has the following characteristics:

  • Centralized job management and scheduling system.
  • Tightly coupled computers.
  • Single system image.

Clusters offer huge potential. The only problem is that, that potential may not be that easy to achieve for most applications. Given the fact that clusters are cheap or free and can be very big and highly available, there are problems why everyone is not using it. One reason is software support. There is very small software support for treating a cluster as a single system.

Another reason is hardware. Network hardware is not intended for parallel processing. Latency is normally very high and bandwidth is relatively low compared to Symmetrical Multiprocessing or SMP and attached processors.

Login/Register to Answer

Related Questions