For most evaluations, running OpenStack on a single node is ideal. It gives you a chance to understand the different cloud services that make up OpenStack, understand how they are configured, and how to troubleshoot errors that you may come across. We've provided an OpenStack Unified Archive to help make that much easier to do - simply modify your Automated Installer manifests to point at this archive, or use the archive to create an Oracle Solaris Kernel Zone.
At some point in time, you'll want to expand this into a multi-node architecture - spreading the load of those cloud services across multiple physical systems. As part of our regular Oracle Solaris 11.2 product documentation set, we've published an Installing and Configuring OpenStack in Oracle Solaris 11.2 document to help you with that. This walks you through a typical small three-node reference architecture that includes a controller node, a network node, and a compute node (conveniently representing the architecture that's also published in the OpenStack upstream documentation).
From this initial three-node setup, it's relatively easy to add more compute nodes, or even split out the cloud storage capabilities into a separate node. This is the first revision of this document, so please give us feedback for how we can improve it!
-- Glynn Foster