MRG Grid FAQ

MRG Grid is the technology that provides the high throughput computing (large amounts of power over a sustained period of time) and high performance computing (large amounts of power over a short period of time) capabilities in Red Hat Enterprise MRG. MRG Grid can schedule tasks onto local grids, remote grids, and harvest idle capacity from desktop workstations. Additionally, Red Hat is working to enable scheduling to rented cloud infrastructures like Amazon EC2 with MRG Grid for the utmost flexibility in providing peak computer power and IT infrastructure utilization.

MRG Grid is based on the Condor Project started and hosted by the University of Wisconsin - Madison. Condor serves as the upstream project for MRG Grid, much like Fedora, kernel.org, and other open source projects serve as the upstream basis for Red Hat Enterprise Linux. And, as with Fedora and other open source projects, Red Hat fully participates in the development community for Condor. MRG Grid builds upon Condor by adding capabilities including enterprise stability and supportability, management tools, and integration with other technologies like MRG Messaging.

Red Hat and the University of Wisconsin announced have signed a strategic partnership whereby the University of Wisconsin will make the Condor source code available under an OSI-approved open source license and Red Hat and the University of Wisconsin will jointly resource and develop Condor on-site at the University of Wisconsin.

For more information, see the MRG Grid Condor page.

MRG Grid enables enterprises to bring large amounts of computing power to workloads, dynamically provision additional peak computing power, and harvest idle capacity from desktop workstations to leverage and utilize existing IT infrastructure. MRG Grid provides both high throughput computing (large amounts of power over a sustained period of time) and high performance computing (large amounts of power over a short period of time) capabilities to process both computationally massive and time critical workloads.

Yes, MRG Grid supports executing on a variety of platforms. Additionally, MRG Grid supports executing in virtual machines so that tasks can run in optimal environments. MRG Grid's scheduler, however, runs only on Red Hat Enterprise Linux.

For a full list of supported platforms, see Red Hat Enterprise MRG Supported Hardware and Platforms.

Red Hat is integrating MRG Grid with the Messaging and Realtime components of Red Hat Enterprise MRG to provide enhanced capabilities. For example, Red Hat is leveraging messaging to add a standardized messaging API from any platform and language to schedule tasks onto a grid that spans from local and idle workstations to remote clouds. Furthermore, MRG's grid capabilities will be able to leverage MRG's large messaging capabilities to transport entire virtual machines efficiently so that machines can execute any task regardless of their host environments. And, Red Hat is working to provide optimized scheduling and improved determinism around grid jobs.