Inspired by http://wiki.daviddarts.com/PirateBox this is for helping with the problem of sparse nodes where they are too far from each other to connect wirelessly/optically. This could be implemented into existing smart phones with the right WiFi support or little boxes like the PirateBox.
I’m calling these mobile store/forward nodes (MSFNs) for non-interactive “push” traffic like e-mail or perhaps a new form of file-sharing. You bring your MSFN to a known location at a given time. Imagine parking this in your car while you shop for groceries or in your backpack while you browse your public library. If you meet indoors with other MSFN administrators, boxes can be plugged into Gigabit switches to greatly increase available bandwidth for exchanges [smartphones need not apply]. If you travel between cities via car, your daily commute could help data propagate across the region of a large city like Los Angeles or Chicago. If you plan on a flight cross-country, you can input an anticipated route and carry a large amount of data for the other coast.
MSFNs advertise prior and anticipated routes, storage capabilities, content and content-requests to all reachable nodes and then start/resume data-transfers (prioritizing the shortest content first, which is typically e-mail). An anticipated route is some path the MSFN will be able to push to in the near future. A prior route might be used for optimizing where MSFNs push data to, if a MSFN often sees a desirable destination is reachable from a given MSFN, it would be better to push in its direction than another.