Abstract
Personal computer performance has increased dramatically over the past decade. So powerful are these machines that they complete most required tasks with time to spare. This spare time is spent in an idle loop, waiting for the next task. A number of organizations have launched public grid computing projects (i.e., using Internet-enabled personal computers) to capture these “lost” CPU cycles via specialized software that effectively aggregates the power of separate systems into a virtual computer. In many respects, computing grids are analogous to electrical power grids (Robb, 2002). Plugging in an electrical power-consuming device taps into the unused capacity of an entire system of power stations and not necessarily the one closest generation station. In computing grids, the individual computers make up a “grid” of processing power stations and the PC owners the associated virtual community.
Key Terms in this Chapter
Amdahl’s Law: Named for computer architect Gene Amdahl, the law describes the theoretical maximum speedup using multiple processors.
Idle Cycles: The repetitive operations performed by a processor while waiting for the start time for the next task.
Grid Operator: The organization that manages the distribution of tasks, collection of results, and day-to-day operations of the computing grid.
SETI@home: A scientific experiment that uses Internet-connected computers in the search for extraterrestrial intelligence (SETI). Participants run a free program that downloads and analyzes radio telescope data for patterns suggestive of intelligent life.
Public Grid Computing: A computing model that provides high throughput computing by using the resources of many separate computers connected the Internet to solve large-scale computation problems.
World Community Grid: Sponsored by IBM, the world community grid’s mission is to create the world’s largest public computing grid to tackle projects involving humanitarian research. The grid will be made available to public and not-for-profit organizations involved in research with broad based benefits to society.
Granularity: Refers to the amount of computation in comparison to the amount of communication necessary to perform a task in distributed task environment