Automated Testing Using Travis CI

Mininet is now using Travis CI for automated testing!

Whenever code is pushed to Mininet, Travis CI starts a test job that installs Mininet, runs a simple sanity test, and the “quick” tests for HiFi, the walkthrough, and the example code.

The full results are available on Travis CI’s Mininet page, and the status of the most recent master build is part of the README on Github:

Travis CI also verifies all new pull requests and displays their build status in the pull request discussion.

Mininet will continue to be tested in several ways on private servers, but our hope is that having a public-facing CI will give members of the community a better sense of how Mininet is tested and the quality of the codebase. Furthermore, we hope that by automatically testing pull requests, we will reduce the likelihood of merging broken code and reduce the testing burden on maintainers.

Announcing Mininet 2.2.1

Mininet 2.2.1 is primarily a performance improvement and bug fix release.

  • Batch startup has been implemented for Open vSwitch, improving startup performance.

  • OVS patch links have been implemented via OVSLink and --link ovs

    Warning! These links have serious limitations compared to virtual Ethernet pairs: they are not attached to real Linux interfaces so you cannot use tcpdump or wireshark with them; they also cannot be used in long chains - we don’t recommend more than 64 OVSLinks, for example --linear,64. However, they can offer significantly better performance than veth pairs, for certain configurations.


  • You can now easily install Mininet on a Raspberry Pi (simply by using -fnv or -a) ;-)

  • mn now accepts multiple --controller arguments

  • --controller remote and RemoteController now accept ip arguments of the form <IP address>:<port> (in addition to ip=<IP>,port=<port>)

Additional information for this release and previous releases may be found in the Release Notes on

The easiest way to get started with Mininet is to download a VM image.

Download, installation, and upgrade instructions may be found at .

The latest Mininet source code is available on github and also via

Thanks to all of the Mininet contributors! A number of interesting features and improvements are already in the works for Mininet 2.3.

Mininet on Raspberry Pi

Mininet is now easy to install on the Raspberry Pi running Debian/Raspbian 7 Wheezy!

Mininet’s has been updated so it should simply work: -fnv  ## basic install -a    ## the works

Here is an example of Mininet running on a Raspberry Pi 2 model B:

A Raspberry Pi can be used any time you need a cheap network-in-a-box, for example for courses, laboratories, demonstrations, or simple experiments.

We look forward to hearing (on mininet-discuss and elsewhere) about interesting, fun and clever uses of Mininet running on the Pi !!

Announcing Mininet 2.2.0 !

We are pleased to announce Mininet 2.2.0! This release provides a number of bug fixes as well as several new features, including:

  • Improved OpenFlow 1.3 support

    • mn --switch ovs,protocols=openflow13 starts OVS in 1.3 mode
    • -w installs 1.3-compatible Wireshark dissector using Loxigen
      • -y installs the Ryu 1.3-compatible controller
  • A new node library, and new Node types including LinuxBridge, OVSBridge, LinuxRouter (see examples/) and NAT

  • Easier LAN/internet connectivity from Mininet hosts

    • mn --nat will semi-automatically connect a Mininet network to your LAN using NAT, solving the “why can’t I ping Google?” problem.

      Warning: By default this will reroute local traffic originating at your Mininet server or VM and destined for Mininet’s IP subnet ( by default) to the Mininet network, which can break connectivity if you are using addresses in the same range in your LAN. You can change this range using the --ipbase option.

  • An improved MiniEdit GUI (examples/ - thanks to Gregory Gee

  • Support for multiple --custom arguments to mn

    mn --custom, --topo mytopo --switch myswitch

  • Experimental cluster support - consult the documentation for details - as well as examples/ and an experimental --cluster option for topologies built with the default Host and OVSSwitch classes:

    mn --cluster localhost,server1,server2,server3 --topo tree,4,4

Note that examples contain experimental features which might “graduate” into mainline Mininet in the future, but they should not be considered a stable part of the Mininet API!

How to get it

The easiest way to get started with Mininet is to download a VM image. See for details.

The latest Mininet source code is available on github and also via

Additional information on the Mininet 2.2.0 release may be found in the release notes and on

Mininet is an open source project, and we are grateful to our many code contributors, bug reporters, and active users who continue to help make Mininet a useful system with a friendly and helpful community!

Have fun with Mininet!

Mininet 2.2.0 Core Team:
Bob Lantz, Brian O’Connor, Cody Burkard
Open Networking Laboratory

Minininet Tutorial at SIGCOMM

Materials from our tutorial at SIGCOMM 2014, “Teaching Computer Networking with Mininet”, are available on The original announcement follows.

Call for Participation:

Mininet Tutorial - Teaching Computer Networking with Mininet

August 18th, Chicago, USA (in conjunction with SIGCOMM)

Important Dates

  • Early Registration Deadline: July 11th, 11:59pm CDT
  • Late Registration: July 12th - August 18th
  • Tutorial Date: August 18th 8:30am - 12:00pm


Sharing our experiences using the Mininet network emulation platform to teach computer networking in a variety of contexts ranging from small on-campus courses (Stanford, Georgia Tech, MIT) to massive online courses. We believe that the Mininet platform can make teaching and learning computer networking more fun and effective by facilitating experiential learning.


  • Introduce Mininet to the broader network community, including its capabilities and limitations and some example use cases.
  • Familiarize attendees with Mininet through hands-on exercises.
  • Share our experiences using Mininet since 2011 for on-campus and massive online courses at Stanford, Georgia Tech and MIT.
  • Disseminate assignments that we used at our universities and help to develop and expand an online repository of Mininet courseware.

Speakers and Panelists

  • Bob Lantz - ON.Lab
  • Te-Yuan (TY) Huang - Stanford
  • Vimalkumar Jeyakumar - Stanford
  • Brian O’Connor - ON.Lab
  • Nick Feamster - Georgia Tech
  • Keith Winstein - MIT
  • Anirudh Sivaraman - MIT

Announcing Mininet 2.1.0 !

We are pleased to announce the immediate availability of Mininet 2.1.0 on !

Mininet is a network emulation platform that is intended to be the quickest, easiest, and most enjoyable way to create virtual networks for research, education, and system development. By providing an instant virtual network on your laptop that can run real application, switch, and controller code, Mininet also makes it convenient to get started with SDN and OpenFlow without any additional hardware.

Mininet 2.1.0 provides a number of bug fixes as well as new features, including:

  • Convenient access to Mininet() as a dict of nodes
  • X11 tunneling (wireshark in Mininet hosts, finally!)
  • Accurate reflection of the Mininet() object in the CLI
  • Automatically detecting and adjusting resource limits
  • Automatic cleanup on failure of the mn command
  • Support for installing the OpenFlow 1.3 versions of the reference user switch and NOX from CPqD and Ericsson
  • Preliminary support for running OVS in user space mode
  • Preliminary support (IVSSwitch()) for the Indigo Virtual Switch
  • Preliminary support for Fedora in
  • A script (util/m) for easily running new shells or commands in nodes
  • The ability to import modules from mininet.examples

We have provided several new examples (which can easily be imported to provide useful functionality) including:

  • Modeling separate control and data networks: mininet.examples.controlnet
  • Connecting Mininet hosts to the internet (or a LAN) using NAT: mininet.examples.nat
  • Creating per-host custom directories using bind mounts: mininet.examples.bind

Note that examples contain experimental features which might “graduate” into mainline Mininet in the future, but they should not be considered a stable part of the Mininet API!

Additional information may be found in the Release Notes.

Mininet is an open source project, and we are grateful to our many code contributors, bug reporters, and active users who continue to help make Mininet a useful system with a friendly and helpful community!

Have fun with Mininet!

The Mininet 2.1.0 Team
Bob Lantz and Brian O’Connor
Open Networking Laboratory

Nick Feamster’s SDN Course

I just want to put in a recommendation for Nick Feamster’s excellent Software Defined Networking course. It’s free on Coursera, and not only does he do a great job of clearly explaining SDN, network virtualization, and related topics, he has also created some excellent videos explaining the usage, design, and rationale for the Mininet system, which is also being used for programming assignments and labs in the course.

There are also interesting interviews with notable people like Martin Casado, David Clark, and Teemu Koponen. It’s quite excellent, very approachable for anyone who has a basic networking background, and highly recommended:

The course has already started, but as of today (Thursday July 11th) you can still sign up for it.


Automating Controller Startup

One of Mininet’s key features is that it makes it very easy to create a complete virtual network including hosts, switches, links and OpenFlow controllers. By default, Mininet runs Open vSwitch in OpenFlow mode, which requires an OpenFlow controller.

As you probably know, Mininet comes with built-in Controller() classes to support several controllers, including the OpenFlow reference controller (controller), Open vSwitch’s ovs-controller, and the now-deprecated NOX Classic.

You can easily choose which one you want when you invoke the mn command:

sudo mn --controller ref
sudo mn --controller ovsc
sudo mn --controller NOX,pyswitch

Each of the above examples uses a controller which turns your OVS switches into Ethernet bridges (aka learning switches.)

Note that ovsc is easy to install but only supports 16 switches. You can install the reference controller (patched to support many switches) using -f. (As of this writing, you can also install NOX Classic using -x.)

The manual way: using remote controllers

But what do you do if you want another controller like POX? Well, one way is to start up POX (or Beacon, Floodlight, etc.) in another terminal window and then use --controller remote:

sudo mn --controller remote,ip=

If your controller is running locally, you can just use

sudo mn --controller remote

You might be asking: is there a way for Mininet to automatically start up my controller so that I don’t have to do it manually? The answer is, of course, yes!

The automatic way #1: using the command line

The latest Mininet master branch allows you to change the command and arguments for --controller ref to invoke another controller as desired. For example, here’s how you could invoke POX:

sudo mn --controller ref,command='python',cdir=$HOME/pox,cargs='openflow.of_01 --port=%s forwarding.l2_learning'

Note the following options:

cdir: the directory to change to before invoking the controller

command: the command to start your controller

cargs: the controller’s arguments (including %s for wherever the controller’s listening port should go)

The super automatic way #2: creating a custom Controller() subclass

This is much easier than it sounds, and it has worked in Mininet since version 1.0!

Here is a custom class for POX’s l2_learning:

from mininet.node import Controller
from os import environ

POXDIR = environ[ 'HOME' ] + '/pox'

class POX( Controller ):
    def __init__( self, name, cdir=POXDIR,
                  cargs=( 'openflow.of_01 --port=%s '
                          'forwarding.l2_learning' ),
                  **kwargs ):
        Controller.__init__( self, name, cdir=cdir,
                             cargs=cargs, **kwargs )

controllers={ 'pox': POX }

With this class saved into, you can now do the following:

sudo mn --custom --controller pox

Additionally you can use the custom class in your Mininet scripts


from mininet.log import setLogLevel
from import Mininet
from mininet.topolib import TreeTopo
from mininet.cli import CLI

# We assume you are in the same directory as
# or that it is loadable via PYTHONPATH
from pox import POX

setLogLevel( 'info' )

net = Mininet( topo=TreeTopo( depth=2, fanout=2 ),
               controller=POX )

CLI( net )

Note that in this custom class we’ve simply changed __init__, but in a more complicated example you might wish to change the start() and stop() methods as well.