Grid Computing

Grids enable the sharing, selection, and aggregation of a wide variety of geographically distributed computational resources (such as supercomputers, compute clusters, storage systems, data sources, instruments, people) and presents them as a single, unified resource for solving large-scale compute and data intensive applications. This idea is analogous to electric power network (grid) where power generators are distributed, but the users are able to access electric power without bothering about the source of energy and its location.

Our lab participates in the Grid4All European Project. Grid4All embraces the vision of a democratic Grid as a ubiquitous utility whereby domestic users, small organisations and enterprises may draw on resources on the Internet without having to individually invest and manage computing and IT resources.

As part of our involvement to Grid4All, we have developed DFS. DFS is our proposal for a data storage substrate. In our design, we clearly separate File and Storage resources. The storage management is provided by the Virtual Block Store (VBS), which is an independent layer, but one that uses the generic interfaces and protocols of the DFS.

DFS Principles

The DFS Architecture is based on the following principles:
  • Each DFS User is identified by a cryptographic certificate. DFS Users are uniquely identified by their cryptographic certificate. With this certificate, they can be named, authenticated, authorized, accounted, attributed an action, associated with an entity or with each other. User throughout this text means DFS User.
  • Any resource is identified by an authority DFS User. Any resource and policy concerning resources is identified by and submitted to the unconditional authority of a unique DFS User.
  • Any action upon a resource has an agent DFS User. The agent of any action concerning resources is ultimately attributed to a single DFS User. This facilitates the accountability of users for their actions and simplifies resource access control.
  • Any action upon a resource is a transaction between an agent user and an authority user. Any resource manipulating action is equivalent to a transaction between the agent of the action and the authority of the resource. The agent must acquire the permission of the authority in order to successfully perform the action.


This topic: CSLab > WebHome > P2P > GRID
Topic revision: r3 - 2008-03-06 - GiorgosVerigakis
 
This site is powered by the TWiki collaboration platform Powered by Perl

No permission to view TWiki.WebBottomBar