There is a standard for syncronizing data between handhelds over GPRS/WAP or internet. It's called SyncML. There is also an opensource server which handels this called
funambol (sync4j). Perhaps you could leverage that? Reinventing a server-client structure + protocol seems like a bad idea? The problem would not be how to store the data, but how to present it in a good accessible manner. Flesh out how it should operate and feel first. Then build an application based on that.
Also, thinking about how to monetize the application also seems like a backward strategy? I suggest first value then profit. I wouldn't use an application that I knew were bulit around a system structure that forces me to pay for no good reason. By that I mean that everybody should have access so that they can set up this system by themselves. But you could provide an service for those who don't want to. That is only my opinion tho.