* Source routing * Dissemination tests
* Mike, JHU * Om, USC * Phil, Stanford * Thanh, PDX
Mike : We have defined the interfaces, allow users to set the route and reset the route.
Phil : why do we need to reset the route?
Mike : Because we don't know how to reach a particular node. Applications leave it to routing, and fowarding engine is very simple.
Om : We should discuss more about what is the good interface for this?
Phil : Pointers might not be efficient, especially when looping through a list and everytime we send a packet, we have to do a loop.
Thanh: Is the route set for every packet being sent?
Om : Yes, for this particular version, there's no routing state so the route is set per packet.
Om : If packets are sent to a destination most of the time, then pointers can be efficient.
Om : What are the possible error cases? incorrect payload length and incorrect path
Mike : It might be interesting to allow intercepting the packets
Thanh : There are multiple potential overlapping testing applications for dissemination protocols. Dhv has been modified so Dhv and Dip
can be tested using the same test application. Also, considering standard test cases for dissemination protocols in general.
Mike : Is this for both TOSSIM and real testbeds
Thanh : Yes
Om : So, we need to define highly representative test cases including test scenarios and network topologies, test applications for
both TOSSIM and real testbeds. This should be described in a separate TEP for benchmarking. This can be useful for collection protocols too.
Mike : Would be nice to include simulated testbeds
Om : Thanh will be leading this direction for now
Thanh : Ok