It's a nightmare. I used to work for one of the industry dinosaurs; they have an MLS team that was as large as their development team. They pull data from over 600 MLSes, with different requirements for each. Managing 600 business contacts is a big pain, then there are different data conventions that change without notice. They have had a long-standing arrangement to feed data to Trulia too.
The word is most MLSes use CDNs, so scraping might not be the best way.
The word is most MLSes use CDNs, so scraping might not be the best way.