(short for 'Virtual Cluster Orchestration System'
) is an environment aware load balancing virtual cluster system
that coordinates virtual machines that offer some service (HPC, SOA, etc) in a failover cluster, making a multipurpose environment which the main goal is to reduce the wasting of the infrastructure's computational potential. ViCOS
applies cloud computing concepts to high performance compute to ensure a flexible, economic and reliable environment for performing tasks demanding high performance.
In order to do that, it uses the available idle machines of the environment, occupying them with virtual machines on demand and adapting to resources changes: if a host is to be used for other purposes, the system migrates the VMs to another host, so the task isn't interrupted and can continue elsewhere.
Therefore, the physical resources are managed without work load interruption, and the system fits itself to the environment, automatically expanding and contracting based on resource demand, focusing in both performance increase and energy savings.
To learn more about ViCOS see the page Documentation
New video of ViCOS in action!
A new video was produced to show how ViCOS works. It presents an environment with ViCOS running, where you can see how the jobs are submitted, and how the VM management occurs.
You can view the video at the documentation
A new version of ViCOS is available!
We released ViCOS 0.2. This version has several changes in the project structure, the components are implemented as Windows services, and we started the support of Linux HPC Clusters. Also, there is a new guide to help you to deploy your own version of the system.
You can view more details about this new version at the ViCOS 0.2
About our team
This project is developed by the Microsoft Innovation Center at Unicamp, in Campinas, Brazil.
You can learn more about our lab at the LMS
page (in portuguese).