Cluster Computing in the Cloud
The elastic nature of cloud computing is interesting to some traditional cluster computing customers who have workloads that have tremendous spikes and valleys. When these people need computing resources, they need a lot of it (hundreds, potentially thousands of cores). But their workloads (often cpu-intensive models) do not need to run all the time (like a webserver, or mailserver).
For these reasons, the pay-as-you-go and elastic nature of cloud computing is interesting.
In some cases, some infrastructure could actually help present resources available in a cloud as a unified system. See:
* Condor
- http://
In other cases, it's a matter of providing parallel programming libraries that are notably missing from Ubuntu. See:
* mpich2
- http://
This blueprint suggests that the Ubuntu server, as guests within a cloud computing system could and should provide the resources needed by cluster computing customers.
:-Dustin
Blueprint information
- Status:
- Complete
- Approver:
- Rick Clark
- Priority:
- Medium
- Drafter:
- Dustin Kirkland
- Direction:
- Approved
- Assignee:
- Dustin Kirkland
- Definition:
- Approved
- Series goal:
- Accepted for karmic
- Implementation:
-
Implemented
- Milestone target:
-
karmic-alpha-5
- Started by
- Dustin Kirkland
- Completed by
- Dustin Kirkland
Whiteboard
= Status =
== MPICH2 ==
* Packaged, in Karmic Universe
== Condor ==
* Packaged, in Karmic Universe
-- Dustin Kirkland
-------
Discussion Points:
* Define difference between cloud and grid
* Why cloud is interesting to grid computing customers?
* Discuss advantages/
* Identify the low hanging fruit
* Some new request-
* Looking a bit further out, what are the higher cost, greater return on investment work items in this space?
-- Dustin Kirkland
* Howto build a virtualized SGE cluster on top of a hybrid cloud using OpenNebula
-- Tino Vazquez
