I am very optimistic about the potential of OpenStack (see previous CloudAve Coverage) and how it can help create open federated cloud ecosystem. Even though OpenStack’s Object Storage is ready for production use, its Compute platform is not. It is bound to change in the first quarter of this year but the code in the Austin release is not ready for production deployment yet. It doesn’t deter progressive organizations like The Wikimedia Foundation from trying out OpenStack.
In a blog post yesterday, Ryan Lane of Wikimedia Foundation talked about how they are testing out OpenStack for their dev & test infrastructure. Along with other technologies, OpenStack will be at the core of their platform, running on 4 nodes (1 controller node and 3 compute nodes). This setup should help them run 100-120 VMs.
OpenStack will play the following roles in their setup:
- Handling VM creation and scheduling on their compute nodes
- Handling IP address allocation
- Gets user account information from LDAP
- Stores IP/VM information in MySQL
PowerDNS is used to handle their DNS needs Puppet is used for their configuration needs and MediaWiki will serve as their VM manager.
This is still a work in progress and Mr. Lane is looking for volunteers to help him out on the project. If you want a hands on experience in the implementation of one of the hottest open source cloud infrastructure technology, please contact him. I am really excited about what Wikimedia folks are doing because this will be a good case study on OpenStack.
- Looking Back 2010: OpenStack Offers Promise (cloudave.com)
- First OpenStack Release Produces Open Source Cloud Alternative (eon.businesswire.com)
- OpenStack Plans Next Two Cloud Platform Releases at First Public Design Summit (eon.businesswire.com)
- OpenStack – an open source cloud platform (enterpriseirregulars.com)
- OpenStack Announces ‘Austin’ Release (datacenterknowledge.com)