ns-3.32 has been used for experimentation. Download the version of ns-3 from : https://www.nsnam.org/releases/ns-3-32/. Copy the contents of the 'scratch' folder to the ns-3 'scratch' folder. For running any experimentation regarding trust, copy the contents of 'src' folder to the ns-3 'src' folder. The contents of source folder are modified versions of Trust Based Routing Protocol presented in https://github.com/JudeNiroshan/ns-3-dev-git/tree/gsoc2018_final. Use the following commands to build and configure ns-3:
python3 ./waf configure --enable-tests --enable-sudo
python3 ./waf build
Following is a tutorial on how to install docker:
https://www.simplilearn.com/tutorials/docker-tutorial/how-to-install-docker-on-ubuntu
Docker network between host device and physical nodes can be achieved using overlay network. For the overlay network, the host device creates a swarm of which physical nodes become parts. In order to initiate the swarm, execute the following command on the host machine:
sudo docker swarm init
The following message will appear:
Swarm initialized: current node (NODE_NAME) is now a manager.
To add a worker to this swarm, run the following command:
docker swarm join --token SOME_TOKEN SOME_IP_ADDRESS
To add a manager to this swarm, run 'docker swarm join-token manager' and follow the instructions.
Now run the following commands for all physical nodes to make them a part of the swarm:
docker swarm join --token SOME_TOKEN SOME_IP_ADDRESS
The following confirmation message will be displayed:
This node joined the swarm as a worker
At this point, the host device and the physical nodes are part of a swarm. For them to communicate, a overlay network has to be created between them. Run the following command on host device terminal:
sudo docker network create -d overlay --attachable NETWORK_NAME
An overlay network has been created named 'NETWORK_NAME'. Two services needs to be created within host device and each physical node using the following command on host machine:
sudo docker service create --name SERVICE_NAME --network NETWORK_NAME --cap-add=NET_ADMIN --replicas 2 alpine sleep 1d
A service named 'SERVICE_NAME' is created in host and one physical device.
We can create a different overlay network and services for a different physical device in the same way. Now the physical nodes can communicate among themselves after going into the shared service of the respective devices using the command:
sudo docker exec -it SERVICE_NAME sh
Repeat these steps for every ghost node connected to ns-3.
In order for ns-3 to communicate with physical devices, three different bridges need to be considered:
- Tap Bridge to bring the data packets out of ns-3 to host machine
Following set of commands create the tap bridge for each physical node:
sudo tunctl -t tap-NAME_OF_TAP_DEVICE
sudo ifconfig tap-NAME_OF_TAP_DEVICE 0.0.0.0 promisc up
- Network Bridge to move the packets inside host machine:
Following command creates a linux bridge for each physical node:
sudo brctl addbr br-NAME_OF_TAP_DEVICE
The following command links the tap bridge to the respective linux bridge:
sudo brctl addif br-NAME_OF_TAP_DEVICE tap-NAME_OF_TAP_DEVICE
sudo ifconfig br-NAME_OF_TAP_DEVICE up
- Bridge to connect linux host to docker container of each host:
Following sets of commands connect each service of host machine linked to a physical node to the respective linux bridge:
PID=$(sudo docker inspect --format '{{ .State.Pid }}' NAME_OF_RESPECTIVE_SERVICE_IN_HOST_MACHINE)
hexchars="0123456789ABCDEF"
end=$( for i in {1..8} ; do echo -n ${hexchars:$(( $RANDOM % 16 )):1} ; done | sed -e 's/\(..\)/:\1/g' )
MAC_ADDR="12:34"$end
sudo mkdir -p /var/run/netns
sudo ln -s /proc/$PID/ns/net /var/run/netns/$PID
sudo ip link add side-int-NAME_OF_TAP_DEVICE type veth peer name side-ext-NAME_OF_TAP_DEVICE
sudo brctl addif br-NAME_OF_TAP_DEVICE side-int-NAME_OF_TAP_DEVICE
sudo ip link set side-int-NAME_OF_TAP_DEVICE up
sudo ip link set side-ext-NAME_OF_TAP_DEVICE netns $PID
sudo ip netns exec $PID ip link set dev side-ext-NAME_OF_TAP_DEVICE name eth3
sudo ip netns exec $PID ip link set eth3 address $MAC_ADDR
sudo ip netns exec $PID ip link set eth3 up
sudo ip netns exec $PID ip addr add IP_ADDRESS_OF_RESPECTIVE_GHOST_NODE_IN_NS3/16 dev eth3
Repeat these steps for every physical node linked to ghost node. At this moment, we can run the ns-3 example codes using the following command:
python3 ./waf --run scratch/tap-wifi-virtual-machine
Static routes have to be set from each of the external node to the simulated nodes inside ns-3. Execute the following command from external node:
ip route add 10.0.1.0/24 via IP_ADDRESS_OF_NEXT_HOP_NODE
Repeat the step for every physical node connected to ns-3.
After experimentation concludes, it is necessary to delete all the tap and network bridges:
sudo ifconfig br-NAME_OF_TAP_DEVICE down
sudo brctl delif br-NAME_OF_TAP_DEVICE tap-NAME_OF_TAP_DEVICE
sudo brctl delbr br-NAME_OF_TAP_DEVICE
sudo ifconfig tap-NAME_OF_TAP_DEVICE down
sudo tunctl -d tap-NAME_OF_TAP_DEVICE
sudo docker service rm SERVICE_NAME
sudo docker network rm NETWORK_NAME
sudo docker swarm leave --force
