Cocktail on the server

We use a lot of third party libraries that are written in .Net. We use rpc to make calls to these libraries or we put them in a windows service and save records in a db to trigger them. On the client I am very confident about how to organize code and use the repository model and unit of work but on the server things are getting a little sloppy.

I used to think that this topic was beyond the scope of cocktail because cocktail is a client side framework and designed to run asynchronously while the server is synchronous. Then I watched John Papa's SPA video and there he was implementing a cocktail like framework on the server and it looked pretty good. I could see Ward Bell/Ideablade influence through the whole video. It seems the Cocktail design principles are gaining traction. John, however,  had implemented it on the server synchronously. 

This brought up questions in my mind again. What is cocktails answer for the server? How should we be organizing our code on the server and in services where there is no need for async but a still prevalent need for organization and code reuses? Why am i confident on the client but not so much on the server?

What are your thoughts?

Thanks for you work on this project Marcel we rely on your advice a lot and it is nice to see people like John Papa creating cocktailesque applications.