In this unit, we learned that a grid is a network of machines that share resources to accomplish larger tasks. The three main resources a grid provides are computation (processing power), storage (memory and hard drives), and communication (network bandwidth). We also explored how expensive software licenses can be managed efficiently by routing jobs only to machines where the software is installed.
We studied how grid jobs are managed through three strategies: scheduling (automatically assigning jobs to suitable machines), reservation (booking resources in advance), and scavenging (using idle machines without disrupting normal users). We also learned that applications can be broken into smaller jobs and subjobs that run in parallel, improving overall performance and efficiency.
Finally, we explored different grid topologies, from small department-level clusters called intragrids to large cross-organization networks called intergrids, where security and policy management become increasingly important. The case studies helped us connect these concepts to real-world situations such as medical imaging, university labs, film rendering, and disaster response systems.