If you mention open-source cloud orchestration tools these days, everyone immediately thinks about OpenStack (including the people who spent months or years trying to make it ready for production use). In the meantime, there are at least two other comparable open-source products (CloudStack and Eucalyptus) that nobody talks about. Obviously having a working product is not as sexy as having 50+ vendors and analysts producing press releases.

Most people who tried to implement OpenStack complain about the complexity (and immaturity) of Neutron, so I was more than interested to talk with Chiradeep Vittal, a CloudStack architect, about the networking aspects of Cloudstack. In our hour-long chat (Episode 9 of Software Gone Wild) we covered (at least) these topics:

• What is CloudStack
• Which hypervisors does it support?
• Can you run multiple hypervisors in the same cloud? Who would ever want to do that?
• How does it implement tenant separation and isolation?
• How does it integrate virtual networks with physical switches?
• How does it connect bare-metal and virtualized servers?
• How would you integrate physical firewalls and load balancers with your CloudStack-orchestrated cloud?
• How can you extend CloudStack and submit code to the project?
• How easy is it to move workload from one CloudStack-orchestrated cloud to another?
• What are StackMate and Apache Brooklyn and where would you use them?
• What are elastic, portable and secondary IPs?
• How do you gradually implement CloudStack in an enterprise environment?
• Can you use CloudStack to manage just a part of your vSphere infrastructure?

Enjoy the podcast and don’t forget to subscribe to the Software Gone Wild feed (Atom, RSS or iTunes).