Building a Global Compute Grid - Two Examples Using the Sun ONE Grid Engine and the Globus Toolkit
- Global Grid Example One
- Overview of a Global Compute Grid
- Brief Description or the Globus Toolkit 2.0
- Distributed Resource Managers
- Portal Software and Authentication
- Global Grid Example One
- Global Grid Example Two
- About the Authors
- About the Authors
- Ordering Sun Documents
- Accessing Sun Documentation Online
Like this article? We recommend
Like this article? We recommend
Like this article? We recommend
This BluePrints article extends the series on compute grids. In the previous BluePrints articles entitled "Introduction to the Cluster Grid" parts 1 and 2, the architecture and design of Sun's cluster grid was discussed (see next section for a description of the cluster grid). In this BluePrints article, the global grid is tackled, first in general terms then by considering two case studies. In particular, we look at the compute grid and the integration of the campus or cluster level grids with global grid technologies.
While this document is technical, it should provide suitable introductory material for those interested in or considering building a global compute grid. References are given at the end of this document to where you can obtain further information about the software elements discussed.
What is a Grid?
Grid software and middleware enhances the way that collections of geographically distributed resources can be shared. Ian Foster, a pioneer in grid computing, describes the grid as "... a flexible, secure, co-ordinated resource sharing among dynamic collections of individuals, institutions and resources.". While the term has traditionally related just to high performance computing, organizations such as Globus and Avaki have helped extend the scope so that the utility of the tools becomes apparent in other fields.
Sun divides grid computing into three logical levels of deployment: cluster grids, enterprise grids, and global grids. global grids are collections of enterprise and cluster grids as well as other geographically distributed resources, all of which have agreed upon global usage policies and protocols to enable resource sharing. Enterprise grids enable multiple projects or departments to share resources within an enterprise or campus and do not have to address the security issues associated with global grids. A cluster grid is the simplest form of a grid and provides a compute service at the local area network (LAN) level. The class of software at the heart of the cluster grid is the distributed resource management (DRM) system.
Furthermore, the term grid can describe different aspects of resource sharing. For example: compute, data, or access. In this BluePrints article, we concentrate on the compute aspect of the grid.