Spread routers to create an interesting network topology for your tests, make sure you can reach all of them through the management network.
|
Note
|
For easier debugging have two separate testbeds: a big one "production" and a small one for debugging. Everytime one protocol finds a bug or some extra configuration are required, the devs will use the small one, without locking the production one. |
|
Note
|
Use cables at least for management network so it just works, Power-line could be useful too. |
-
Connect one of your interfaces to the management network
-
Verify you have proper connectivity to testbed nodes
-
For each routing protocol run the test ex:
cd tests/failure_recovery/; ./test_iperf_fiveflows.sh OLSRv1 -
After the test fish parse the collected data ex:
cd tests/failure_recovery/parser_scripts; ./parse_all_iperf_fiveflows.sh -
Generate the grafs ex:
cd tests/failure_recovery/gnuplot; gnuplot plot_iperf_fiveflows.gnuplot -
Contemplate the beaty in
tests/failure_recovery/graphs/