Home

GRID COMPUTING

Contents

Introduction

ClusterVision is dedicated to the development, implementation and utilisation of Grid Computing technology.ClusterVision supplies Grid-enabled clusters, implements Virtual Laboratories and provides consultancy to organisations who want to benefit from using Grid technology within their organisation.

ClusterVision works closely with the Research Institute for Computer Science from the University of Amsterdam (UvA). The Research Institute for Computer Science has been at the forefront of Grid Computing technology since the early days. The Institute is internationally renowned for its research and software in the field of Grid Computing in general and Virtual Laboratory technology in particular.

Grid Computing Consultancy

ClusterVision provides consultancy to commercial, governmental and educational organisations on the possibilities, advantages and implementation of Grid technology within their organisation.

Our specialists in Grid Computing can advise you on how Grid Technology can benefit your organisation and on what needs to be implemented to gain those benefits. Our expertise in and experience with Grid Technology, High Performance Computing, Grid software, and scientific and engineering applications enables us to give a balanced advice.

We have experience with and can implement a range of both open source and proprietary Grid Software, but are not exclusively bound to one software vendor. Therefore, we can give independent advice and recommend the best solution based on suitability, price and performance.

Grid-Enabling of Clusters

ClusterVision can "grid-enable" your cluster and integrate them in your own Grid infrastructure or in inter-organisational or international Grids.

The grid-enabling of a cluster usually entails at least the following:

  • Evaluation of the customer's wishes and requirements with regards to Grid Computing.
  • Installation and configuration of the Globus Toolkit.
  • Configuration of the Globus interface with the cluster's queuing and scheduling system.
  • Application and implementation of the Globus authentication certificate;
  • Testing the Grid implementation.

Globus is the most popular Grid middleware used on ClusterVision's clusters, but other software can be implemented instead of or in addition to Globus, including Condor-G and Sun Grid Engine.

ClusterVision has supplied the clusters for many Grids, including:

  • UK National Grid Service (NGS)
  • DAS-3 Grid
  • AlmaGrid
  • Grid5000

What is Grid Computing

Grid middleware provides a software layer that allows optimal use of geographically distributed compute, storage and other networked resources in a transparent way.

The users of a Grid no longer have to submit computational jobs to individual computers, but to the Grid as a whole. The Grid then takes care that the jobs gets submitted to the computer(s) that can best process the job in terms of capability, capacity, speed or cost.

The same principal holds for the request of storage capacity or network bandwidth, but also for requests to use certain research equipment such as an MRI scanner or electron microscope.

Grid Computing offers two main advantages to organisations:

1) Grid technology allows scientists and engineers to utilise larger amounts of computer resources together to solve computational problems quicker than would otherwise be possible.

2) Grid technology allows organisations to utilise their computer infrastructure more efficiently by allowing sophisticated software rather than humans to allocate, schedule and account for resources.

 

Virtual Laboratory

One of ClusterVision's services is the implementation of so-called Virtual Laboratories (VL). A Virtual Laboratory is a co-operation between different organisations and people for the distributed analysis of applied scientific experiments, using the internet, Grid Computing technology and VL software. ClusterVision works together with the VL-e project (Virtual Laboratory for e-Science).

A Virtual Laboratory implementation has the following elements:

  • A specialised upper middleware to bridge the gap between the applications and grid-services layer;
  • Unique federated information management provisions supporting and modelling experiments;
  • Means for resource integration and collaboration.

 
Copyright 2002-2012 ClusterVision BV