Summary of Unit 2: Grid Terms and Concepts
A grid is a collection of linked devices, sometimes referred to as nodes or resources, that cooperate to share storage, software, computing power, and communication capabilities in order to accomplish tasks more quickly
Computation (CPU processing power) is a grid's most valuable resource. There are three primary ways that grid systems might use processors: performing the same program repeatedly across several machines, splitting applications into smaller components that operate in parallel, and running applications on distant machines. Scalability, which evaluates how well performance increases with the addition of more processors, is a crucial idea in computation.
Storage is another important resource. A data grid is a grid that incorporates distributed storage. Computers provide secondary storage, including hard drives and memory. Data can be shared across several computers under a single, unified system thanks to distributed file systems like the Network File System and General Parallel File System. By distributing data across several storage devices, techniques like data striping enhance performance by boosting transfer speed and dependability.
An essential grid resource is communication as well. Jobs and the data they require arrive at the appropriate processing nodes when data is transferred across machines efficiently. High-speed and redundant network connections are frequently required to prevent bottlenecks and increase reliability since limited capacity can impair performance.
By only sending jobs to computers with the necessary software installed, grid systems can also handle pricey software and licenses. The licensed limit is not exceeded by the permitted number of users thanks to license management. Furthermore, while allocating tasks, the architecture, operating systems, and rules of the computers in a grid may vary.
In a grid, work is arranged into jobs and applications. One or more jobs that are planned to run on the grid make up an application. Workflow structure can be formed by jobs that depend on one another or execute in parallel. After all connected jobs are finished and their outcomes are merged, the final output is created.
Additionally, scheduling, reservations, and scavenging are explained in the unit. Based on available resources, scheduling automatically allocates jobs to the right machines. Reservations make it possible to reserve resources ahead of time, ensuring deadlines are met. Scavenging use idle computers to carry out tasks without interfering with their principal users, albeit this could result in erratic completion times.
Lastly, the unit makes a distinction between various grid configurations. One department's modest, uniform grid could look like a cluster. Whereas an intergrid links various organizations for resource sharing and collaboration, an intragrid links several departments inside a single organization. Security, rules, and administration become more crucial as grids grow.
In conclusion, this unit provides a clear understanding of how grid systems operate, how resources are shared and managed, and how different components interact to deliver efficient and scalable computing services.