Reflection Unit 2: Grid Terms and Concepts
COE4261 – Grid Computing
In this unit, I learned that grid computing is a structured and managed way of combining distributed machines to share resources efficiently. What I understood most clearly is that a grid is not just hardware connected by a network; it requires management software, scheduling systems, monitoring tools, and security mechanisms to function properly.
One key idea I understood is how jobs and applications operate within a grid. Applications can be divided into multiple jobs and subjobs, and some may run in parallel while others depend on the completion of previous tasks. This showed me that workflow management is very important in grid environments. I also learned that communication between subjobs, often through Message Passing Interface (MPI), is necessary for some scientific or high-performance applications.
The role of donor software also became clear to me. Each machine that contributes resources must authenticate itself, monitor its load, and accept jobs from the grid management system. This ensures secure and controlled participation. I also understood how load sensors and measurement tools help the scheduler make intelligent decisions based on real-time resource usage.
Another concept that stood out to me was the difference between intragrid and intergrid systems. As grids expand across departments or organizations, security and policy enforcement become more critical. This means authentication, authorization, and resource restrictions must be carefully managed.
In summary, this unit helped me see grid computing as a coordinated ecosystem where computing power, storage, communication, policies, and management components work together. It deepened my understanding of distributed systems and the challenges involved in optimizing performance while maintaining security and reliability.