The primary options would be IS-IS and OSPF. I chose IS-IS with our deployment primarily because it scales better. Sure you can break up OSPF into different areas when your network gets really big however with IS-IS you don't need to worry about that. Also IS-IS re-convergence is faster than OSPF due to how it handles neighbor changes and sends out updates (also the reason why it scales better but I feel these differences are worth distinguishing).
At the end of the day IS-IS is the suggested IGP for SD-A but you are not required to use it. If you are more comfortable with OSPF then by all means use OSPF (or other supported underlay protocols). Also, if you feel your network isn't going to scale to a very large amount of NADs then OSPF is a good choice.
I hope this is helpful and I'm happy to answer any follow up questions you may have,
Chuck McFadden
--Please mark helpful posts as helpful--
