We should define test scenarios with the focus on limiting to a bare minimum and simultanously have all possible use cases covered.
I can imagine this:
static large network dense
static large network full
Where dense means a router has in average two, three direct neighbors. Full means 5, 10, 20 neighbors. Probably this is a parameter and the distance and the area is modifyable.
The number of routers within should be a parameter.
This topology is suited to analyse the routing message size for different topologies.
Based on this we can then analyse how the partial update performs. Right, for this static Scenarios the partial mode should perform superior.
Sidenote: we should design in the protocol a automatic that after n partial messages a full message SHOULD be send.
Based on this scenario we should add somehow dynamic characteristics
nodes disappear and appear. The number and random interval of disappearing and appearing nodes should be configurable
We should define test scenarios with the focus on limiting to a bare minimum and simultanously have all possible use cases covered.
I can imagine this:
static large network dense static large network full
Where dense means a router has in average two, three direct neighbors. Full means 5, 10, 20 neighbors. Probably this is a parameter and the distance and the area is modifyable.
The number of routers within should be a parameter.
This topology is suited to analyse the routing message size for different topologies.
Based on this we can then analyse how the partial update performs. Right, for this static Scenarios the partial mode should perform superior.
Sidenote: we should design in the protocol a automatic that after n partial messages a full message SHOULD be send.
Based on this scenario we should add somehow dynamic characteristics