The tool is an implementation of the idea proposed in the article

Kim, Hwangnam, Hyuk Lim, and Jennifer C. Hou. Network Invariant-Based Fast Simulation for Large Scale TCP/IP Networks.

but using fast prototyping (based on Mininet) rather than simulation, which helps to improve accurancy of the method.

1. Install Mininet or use provided vm-image.
2. For result visualization we use bowtie, so one needs to install it prior running the visualization module.

Define network configuration in config.py.

• scale_factor is a number from 0 to 1
• save_path is a folder where tcpdumps transformed into csv-files will be stored
• plot_graph set as True if before running the experiment you want to have a visialization of your topology
• hosts list of hosts
• switches list of switches
• links list of links with theirs parameters (Bandwidth, delay)
• commands field is what will be running on the given host after sleep time and it is supposed to run network application. User may want to use the apps from our library NetworkApps (like FTP server, FTP client, HTTP-server, HTTP-client), built-in linux apps (like AB, curl, etc.) or use traffic generator (D-ITG is the best fit)

There is two options to use it.

python WebInterface.py build
python WebInterface.py serve

This will run a server, available on https://localhost:9991 The experiment can be ran by apploading a config.

Although, if at least one of the applications doesn't finish (usually, it happends to servers), the experiment will be frozen and the results won't show off. In this case, you can interrupt the execution any time. pcap files still will be available in csv folder, and can be analyzed by wireshark, but the user must take into account the fact that it was scaled up by scale factor, which made all the events delayed.

For standalone experiment result (csv) visualization, that will take care of

python WebInterface.py build
python WebInterface.py serve

1. Run the app

sudo python main.py

and wait until the emulation finishes.

2. Folder csv will contain csv- and pcap-files from

Visualize the result

python WebInterface.py build
python WebInterface.py serve

and the graphs will be avaliable via the link https://localhost:9991