In Unit 2, I have learned basic grid computing concepts and how different resources are shared and regulated in a grid system.
A grid is a group of interconnected computers that share their computation, storage, communication capacity, software, and specialized hardware. The machines can have different architectures and operating systems, but they work together to perform large-scale or intricate operations efficiently.
The main types of resources in a grid are;
Computation (CPU power): They can be run on remote machines, broken into parallel tasks, or run multiple times on different nodes. Scalability, I learned is the measure of how well extra processors reduce execution time.
Storage: Grids aggregate storage from many machines using distributed file systems. Journaling increases reliability and data recovery, while data striping offers improved performance.
Communication: Network bandwidth is necessary since jobs may need to transmit data among nodes. If there is limited bandwidth, grid performance will be reduced.
Software and licenses: Some machines contain expensive licensed software; therefore, the scheduler sends particular jobs to only machines that contain the relevant software.
Finally, I learned that the grid work is divided into applications, jobs, and subjobs. Some jobs can run in parallel while some are dependent on the output of the previous job.
Another important concept is resource management through;
Scheduling, where jobs are automatically allocated to appropriate machines.
Reservation – booking resources ahead of time.
Scavenging, as discussed above, is a technique where unused CPU cycles of idle machines are temporarily offered.
Finally, I learned the difference between clusters, intragrids, and intergrids as well as the security, policy, and management complexities involved in spreading grids across departments or organizations.
Overall, this unit helped me to understand how distributed resources are coordinated efficiently, securely and intelligently in a grid computing environment.