You bet it is!

However, common sense has always told me that an index should provide a fast way of locating a record it points to (or otherwise it would become unmaintainable), and storing the record's pointer along with the index keys (be it RID or a clustered key) is the best way to achieve this.

I know for certain that InnoDB does it, but InnoDB tables, being B-Trees rather than B+Trees, do not maintain direct links between leaf-level pages (and do not even distinguish between key-level and leaf-level pages).

I was just seeking for a proof that SQL Server does the same, and, thanks to you, I got it now 🙂