Both clustered indexes and non-clustered indexes are ordered.
There are two advantages with a covering non-clustered index, both coming from the fact that all data that is needed can be read from the NC index alone:
- No need for key lookups. If you need read many rows the key lookups can be expensive.
- In case of a scan, the NC index is typically cheaper to scan, because it is smaller. (The exception is if the NC index includes all columns in the table, in which case it's on par with the clustered index.)
I guess that by "cavitate", you mean caveat. Yes, there is the general caveat: they add overheads to updates.