First, there is a big distinction between need to have and should have. The first list is fairly small whereas the second is not. To answer your specific bullet points, you don't strictly *need* to have either of those things, but they fall into the "should have" category. Here's some starter information for both of those lists.
| Need to have | Should have |
|---|
| vCenter version compatible with target version of vRA | Distributed switch |
Dedicated service account with appropriate permissions inside vCenter | Datastore clusters in fully-automated mode |
| VM templates (if using clone workflows) | Dedicated VM templates with customization specs. |
| Valid license for vRA | Consistent networking and storage configuration |
| Necessary servers for MS SQL and Windows IaaS role(s). | Stable and reliable vSphere with vCenter infrastructure |
But more important than either of these lists is a set of processes and operational methods which are consistent and repeatable in your environment. vRA enables cloud--not simply just convenient VM deployment. If you aren't ready for cloud due to bad process and non-automate-able tasks, you're going to have a hard go of vRA.
