We've all been there. Multiple users working on a central model coordinating who can sync when, either through the work-sharing monitor or a parallel group lync chat.
This has always seemed super inefficient to me.
My idea for C4R would be a stacking queue, so that while one sync is taking place you can add your sync to the queue so that yours will automatically be processed once the central model is available.
With this in place its not an un-imaginable leap that by knowing how may syncs are in the queue and volume of data that need to be transferred that an estimated waiting time could be given allowing users to plan their work and syncing more effectively.