• The single biggest thing I would do is allow the option to select not only which distributor to use, but which distribution database. Replication shouldn't be bound at a server/server level. I support a high-volume replication environment and the biggest pain point is contention in the distribution database. When one publication decides to update 400+ million rows and sends that transaction over to the distributor, everything else comes crashing down with it. (For those that would argue against letting something update 400+ million rows at a time, my response is: I agree, but the API won't let us do that. 🙂 )

    Let us select not only a distribution server, but also a distribution database, even if it's on the same dist server.