I used to be involved with this sort of thing, and it's actually pretty interesting stuff. There are enormous open source projects, operating largely out of public sight, that have recreated the WoW server backend in most of its glory. The lower-level content works quite well these days, though a lot of complex raid scripts are missing or broken. This one is probably the most popular: https://github.com/TrinityCore/T…
One could argue that the Lich King was the last "good" patch for WoW. What I've never understood though was there is clearly some residual demand for "down rev" versions of the game where the mechanics favored a more engaged play style, why not facilitate it? How much does it really cost to run a single realm at the Vanilla, TBC, and LK levels? So many new players where their first experience with Onxyia was soloing…
I don't know exactly how well it's working out for them. Now they have two games to support and promote, with about the same number of users. This number might (likely?) be higher than if they had stuck to providing a single game, but I'm not sure. It's hard to tell.
There seems to be some business risk to doing this. Apart from possibly cannibalizing your main product, it also requires two teams to support and run the two games. If you actually want the experience on the old-style server to be at all stable or compelling, you'll have to put some time and energy into them—it's more than just running the server.
Honestly, I have no idea if what RuneScape did was the right move, but it's a really interesting case study.