On Mon, 2008-12-08 at 12:25 -0800, Javier Cardona wrote: > > That's something we can't actually simulate since you're trying to see > > whether they do NAV updates correctly, right? > > I was trying to simulate hidden nodes, which are unavoidable given the > distributed nature of the mesh. > Also trying to simulate "starvation at the edges", which varies > greatly with the collision avoidance threshold. > But, hey, I was just day dreaming. Even the simpler air groups that > you proposed would be very useful. :) Makes a lot of sense, and you can maybe even implement that when a userspace program is doing the packet propagation, but for topology testing I think we should start with the simpler groups thing. And as I said, the userspace propagation will actually need to turn off in-kernel propagation which fits setting airgroups to 0. johannes
Attachment:
signature.asc
Description: This is a digitally signed message part
