Because it doesn’t require a common cluster, a distributed availability group allows us to link servers in situations were a cluster isn’t possible. Servers could be in remote locations, members or different domains, different OS levels or even different operating systems (yes, we can link Windows and Linux based AGs).
The ability to link servers in this way gives us a very nice and easy way to replicate data between servers when thinking about a migration.
David’s scenario involves a SQL Server upgrade. I’ve seen this work really well in practice for a 2017 to 2019 upgrade. With applications pointing to the listener agent and everything in place, you can have as little as a few seconds of downtime for that upgrade, which is amazing when you think about how we’ve historically migrated to new versions of SQL Server.