Grid computing is the application of several computers to a single problem at the same time – usually to a scientific or technical problem that requires a great number of computer processing cycles or access to large amounts of data. Computational grids that couple geographically distributed resources are becoming the effective computing platform for solving large-scale problems in science, engineering, and commerce. According to John Patrick, IBM’s vice president for Internet strategies, “the next big thing will be grid computing.” Although Grid computing is firmly ensconced in the realm of academic and research activities, more and more companies are starting to turn to it for solving hard-nosed, real-world problems.
Grid computing is emerging as a viable technology that businesses can use to wring more profits and productivity out of IT resources –and it’s going to be up to you developers and administrators to understand Grid computing and put it to work. It’s really more about bringing a problem to the computer (or Grid) and getting a solution to that problem. Grid computing is flexible, secure, coordinated resource sharing among dynamic collections of individuals, institutions, and resources. Grid computing enables the virtualization of distributed computing resources such as processing, network bandwidth, and storage capacity to create a single system image, granting users and applications seamless access to vast IT capabilities. Just as an Internet user views a unified instance of content via the World Wide Web, a Grid user essentially sees a single, large, virtual computer. Grid computing will give worldwide access to a network of distributed resources CPU cycles, storage capacity, devices for input and output, services, whole applications, and more abstract elements like licenses and certificates. For example, to solve a compute-intensive problem, the problem is split into multiple tasks that are distributed over local and remote systems, and the individual results are consolidated at the end. Viewed from another perspective, these systems are connected to one big computing Grid. The individual nodes can have different architectures, operating systems, and software versions. Some of the target systems can be clusters of nodes themselves or high performance servers.
Grid computing can be divided into three logical levels of deployment: Cluster Grids, Enterprise Grids, and Global Grids.
Cluster Grids
The simplest form of a grid, a Cluster Grid consists of multiple systems interconnected through a network. Cluster Grids may contain distributed workstations and servers, as well as centralized resources in a datacenter environment. Typically owned and used by a single project or department, Cluster Grids support both high throughput and high performance jobs. Common examples of the Cluster Grid architecture include compute farms, groups of multi-processor HPC systems, Beowulf clusters, and networks of workstations (NOW).
Enterprise Grids
As capacity needs increase, multiple Cluster Grids can be combined into an Enterprise Grid. Enterprise Grids enable multiple projects or departments to share computing resources in a cooperative way. Enterprise Grids typically contain resources from multiple administrative domains, but are located in the same geographic location.
Global Grids
Global Grids are a collection of Enterprise Grids, all of which have agreed upon global usage policies and protocols, but not necessarily the same implementation. Computing resources may be geographically dispersed, connecting sites around the globe. Designed to support and address the needs of multiple sites and organizations sharing resources, Global Grids provide the power of distributed resources to users anywhere in the world.
