Please add helpful Frequently Asked Questions (FAQs) and high-quality answers (HQAs) below. Before you send a question to mininet-discuss, make sure your question isn't already in the FAQ - and if you see a question asked repeatedly, feel free to add the answer to this FAQ! Since github wiki markdown doesn't automatically generate a Table of Contents, please follow the existing format and add 1) an anchor for your question and 2) a link to it at the top of the page.

• Mn Topo For Mac S
• Mn Topo For Mac Free

Start Here. VM and Installation Questions.

Using Mininet. OpenFlow Questions. Troubleshooting. Advice for the Desperate and/or Clueless!. How can I get started with Mininet?

Quick answer: Follow the steps on our page! The best way to get started with Mininet is to install our ready-to-run virtual machine image as per our instructions, then go through the, and then continue with the other steps on our page.

You may find the and the to be particularly useful in helping you to understand and use Mininet and OpenFlow/Software-Defined Networking, respectively. What is the login/password for the Mininet VM? As of Mininet 2.0.0 and newer, it is currently: mininet-vm login: mininet Password: mininet On some older VMs it was openflow/ openflow. Why can't I get X11 forwarding to work? I get cannot open display: or $DISPLAY not set, and wireshark doesn't work! Xterm doesn't work either! This is not a Mininet problem.

It means that X11 forwarding is not set up correctly. First, consult the X11 setup instructions in the, including:. Download X11. Install X11. Start up X11 server.

Access VM via ssh Make sure you have carefully followed the necessary steps. If things are still not working for you, you will want to make sure that:. Your X11 server on your client machine (e.g.

Your laptop) is installed correctly and is actually running. You are connecting via ssh using X11 forwarding (e.g. Ssh -X on OS X/Linux or enabling X11 forwarding in a Windows ssh client like PuTTY or SecureCRT.). You don't have any options in your client.ssh/config which interfere with X11 forwarding When you log in with ssh, your $DISPLAY environment variable is set X11 terminology is a bit confusing because the X11 server is actually run on the ssh client machine! The ssh client connects to the sshd server, which in turn forwards connections from X11 client applications (such as wireshark) to the local X11 server (usually running on your laptop or whatever machine is sitting in front of you.) One note: if you have disabled IPv6, you may find that you need to add AddressFamily inet to your /etc/ssh/sshdconfig. Other unlikely causes of breaking X11 forwarding include it being disabled in /etc/sshdconfig or disabled by SELinux. Neither of these should be the case in the Mininet VM image we provide.

You may wish to invoke debug logging on your ssh client to see why X11 forwarding isn't working. On OS X and Linux, you can use ssh -X -v to see where the X11 forwarding is failing. By default ssh -X times out after a while - you may prefer ssh -Y for that reason, although it is less secure. There is a wealth of information on the internet explaining how to set up X11 forwarding correctly on any platform.

This is easily found using Google or the search engine of your choice. As an alternative to X11, you could also use VNC, but that is probably about as complicated as X11 and is left as an exercise to the reader. If this seems too complicated, you can simply run X11 in the VM console window as described below! X11 forwarding is too hard! Can't I just run a GUI in my VM console window? Yes, you can!

First, log in to the VM in its console window (i.e. Type directly into the VM window without using ssh) and make sure apt is up to date: sudo apt-get update Then, install the desktop environment of your choice. Sudo apt-get install xinit where is your GUI environment of choice. Some options:.

lxde: a reasonably compact and and fast desktop environment. flwm: a smaller but more primitive desktop environment. ubuntu-desktop: the full, heavyweight Ubuntu Unity desktop environment Then, you can start X11 in the VM console window using startx If you are running VirtualBox, you will want to install the VirtualBox Guest Additions using sudo apt-get install virtualbox-guest-dkms Reboot the VM, log in and run startx, and you should be able to resize the VM console window and desktop. Can I run a GUI/X11 application within a Mininet host? Yes, you can do so from a host xterm with the current version of Mininet.

This allows you to run programs like wireshark or firefox. You might want to su to another user (e.g. Mininet) to avoid running FireFox with root privileges, but it probably doesn't make a fundamental difference in a Mininet VM that is configured for password-less sudo.

The CLI xterm command actually sets up an X11 tunnel which you can continue to use, e.g. Mininet xterm h1 mininet h1 wireshark & You can also use the x command to set up the X11 tunnel (and optionally run an X program): mininet x h1 xclock & mininet x h2 mininet h2 wireshark & What about Mininet 2.0.0? (Thanks to Murphy McCauley for providing the following workaround, which craftily uses the switch's CPU port!) This is a bit of a hack at this moment but it works! Using a graphical browser requires that you get X11 traffic out of your Mininet host namespace and into the environment where you actually have an X display. Say the X display is the host environment with IP address of 192.168.56.1 and the Mininet VM has an IP of 192.168.56.101. In short, run sshd inside Mininet's h1.

Then SSH from the host environment (192.168.56.1) to the Mininet VM (192.168.56.101) with X forwarding (ssh -Y mininet@192.168.56.101). Then SSH from the Mininet VM into h1 with X forwarding (again!). As an example, open three terminals in the host environment (Term1, Term2, Term3) On Term1:./pox.py forwarding.l2learning # Run an OpenFlow controller On Term2: ssh -Y mininet@192.168.56.101 # SSH into the Mininet VM with X forwarding sudo mn -topo=linear,2 -mac -controller=remote,ip=192.168.56.1:6633 h1 /usr/sbin/sshd # From the mininet prompt, run sshd inside the h1 namespace On Term3: ssh -Y mininet@192.168.56.101 # SSH into the Mininet VM with X forwarding sudo ifconfig s1 10.12.12.12 # Give the internal adapter for s1 an address #By default, all hosts live on 10.0.0.0/8 space. Ssh -Y mininet@10.0.0.1 # SSH into the Mininet h1 namespace with X forwarding xeyes # Run any X app How do I run Linux programs on my Mininet hosts? If you are asking this question, it means you haven't yet consulted the.:( If you are incredibly lazy, please at least look at the.

Mn Topo For Mac S

How do I figure out the command-line options for the mn command? Mn -help How can I install Mininet natively on my Linux machine? Instructions for native installation can be found at and in. How can I uninstall Mininet? If you installed Mininet using apt-get install mininet, you can uninstall it using: apt-get remove mininet If you installed from source, there isn’t currently an automatic way to uninstall it. (If someone would like to add reliable, verified uninstall target and/or install.sh option, we’d welcome a pull request!) In the mean time, you might try something like: sudo pip uninstall mininet sudo rm `which mn` sudo rm `which mnexec` sudo rm /usr/share/man/man1/mn.1.

sudo rm /usr/share/man/man1/mnexec.1. lather/rinse/repeat if you have multiple Mininet packages installed.

Note that this procedure will simply uninstall Mininet itself - it will not remove Open vSwitch, the Stanford reference switch or controller, or any other related software which may be installed on your system. I can't boot my VM in VirtualBox in Windows! Do I need a 32-bit VM?

If you are already running Microsoft's Hyper-V, you may not be able to boot the 64-bit Mininet VM in VirtualBox at the same time. I tested this and got the following error: VirtualBox - Error VT-X/AMD-V Hardware acceleration is not available on your system. Your 64-bit guest will fail to detect a 64-bit CPU and will not be able to boot. Usually the problem is not that you don't have a 64-bit CPU (you probably do if you have anything as good as, say an intel Core 2 Duo from 2006.) It is much more likely that:. You need to enable VT-X/AMD hardware virtualization in the BIOS, or:. You are trying to run two virtual machine monitors - Hyper-V and VirtualBox - at the same time, and this does not work with 64-bit guest OSes. A simple solution to this problem is to turn off Hyper-V (in Windows 8, this is done via 'Enable/Disable Windows Features' in the Windows Control Panel.) This has been tested and verified on Windows 8.

If you need to keep running Hyper-V, you could try the 32-bit Mininet VM image, or you could run the Mininet VM image natively in Hyper-V by:. Converting the.vmdk disk image to a.vhd using Microsoft's Virtual Machine Converter(. Creating a new Hyper-V virtual machine using the new.vhd image as its hard drive. Creating an 'external' virtual switch in Hyper-V manager, sharing the interface with the host OS. I have a netbook from 2005 and I really want a 32-bit VM (or maybe I need one for testing or for nested virtualization on older hardware) OK, use the 32-bit VM image.;-) Help! The VM console screen is blank!

Make sure that the VM is actually booting without any errors of any kind. If not, then you may have a problem actually booting the Mininet VM in your VM monitor. The 64-bit Mininet VM image should work on any modern CPU.

However, some users have reported conflicts between Microsoft's Hyper-V and VirtualBox on Windows, so if you are running VirtualBox on Windows you may need to turn off Hyper-V. Alternately, you can use Hyper-V to run the VM as noted above! Or you can try the 32-bit image, which seems to work fine on most configurations. If the VM is booting but boots to a blank screen, then you probably just need to either wake up the Linux console or switch to another virtual console, as follows: First, select the VM console window. Second, press a key like A or return a few times - see if any text appears. If nothing happens, try switching to a different Linux console using control``alt``F1`` through control altF7`. Note: On a Mac laptop using VMware Fusion, you may need to type fn``control``option``F1, since the F1 key controls brightness by default.

I can't import the.ovf into VirtualBox/VMware/etc.! Unfortunately, when VirtualBox was updated to 4.3.4, it stopped being able to import some versions of the Mininet.ovf file. Some other VMMs may also not be able to import it directly, but there is an easy workaround: If you extract the.zip file, you should see a.vmdk disk image file. You should be able to create a new virtual machine in VirtualBox/VMware/etc. and you should be able to specify that this new VM should use an existing disk image file, and you should select the Mininet.vmdk file. Configure, boot, and enjoy! I can't connect to the internet on the Mininet VM to install packages!

(or, I can't SSH into the mininet VM from the host box) In VirtualBox, you need two different network interfaces set up if you want to both access the internet from your VM and access your VM from the host. One of them should be a NAT interface (to get to the internet), and the other should be a host-only interface (to get to, well, the host). Set up the interfaces in VirtualBox that way, then add eth0 and eth1 lines in the VM's /etc/network/interfaces as below: # The host-only interface auto eth0 iface eth0 inet dhcp # The internet interface auto eth1 iface eth1 inet dhcp How do I use Mininet's Python API?

You are asking the right question!! The Python API open's up Mininet's full potential. Check out the for an introduction to Mininet and its Python API. Several useful examples of using the Python API can also be found in the directory. We also provide Python DocStrings for every Mininet class and method, and you can view them using from within Python import mininet.node help(mininet.node.Node) or by calling Python from within the Mininet CLI: mininet py help(h2) In each case, pressing q should quit the pager.

The API documentation is also available at. Also see for information on how to generate Mininet documentation yourself in.html and.pdf format. Why can't I ping Google from my Mininet hosts? You can't ping google.com because your Mininet network is not connected to the internet. This is usually a good thing!

Usually Mininet networks use a non-routable IP address range like 10.0.0.0/8. However, you can set up NAT if you like. How can I set up NAT? In Mininet 2.2 and newer, you can use the -nat option: mn -nat. In order for DNS to work in the Mininet hosts, you should not be using dnsmasq for local caching.

Switch = partial( OVSSwitch, protocols = 'OpenFlow13 ' ) net = Mininet( topo, switch =switch. ) Open vSwitch 2.3 and newer support 1.3 by default. It is easy to install it using: install.sh -V 2.3.1 What OpenFlow switch implementations does Mininet support? Mininet currently includes support for the user space reference implementations, Open vSwitch in kernel and user space modes, and the Indigo Virtual Switch.

The reference switch and OVS are included in the VM image, and IVS can easily be installed using install.sh -i. Mininet used to support the OpenFlow 0.8.9 kernel reference implementation ( -switch kernel) but that is now obsolete and has largely been replaced with Open vSwitch. The command line options are -switch user and -switch ovsk for the user reference and Open vSwitch kernel switches, respectively. You can also install the CPqD switch using install.sh -3f; it will replace the Stanford reference switch, i.e.switch user and UserSwitch.

See below for an example of using it. Why does my controller, which implements an Ethernet bridge or learning switch, not work with my network which has loops in it? I can't ping anything! Tl;dr: use -switch lxbr,stp=1 or -switch ovsbr,stp=1 and wait for STP to converge. It doesn't work because your network has loops in it. Transparent bridging of L2/Ethernet networks doesn't work if the topology has loops in it, for a variety of reasons: ARP does broadcasts, packets are flooded by default, learning switches don't deal well with seeing the same MAC address on multiple ports and could potentially learn a route to themselves, and Ethernet frames don't have a time to live field (TTL) the way IP packets do (otherwise flooding might work, if inefficiently.) As a result, many Ethernet bridges implement variants of a Spanning Tree Protocol (STP), which simply deactivates links in the network to remove loops.

Of course, this also throws away network bandwidth that you could otherwise be using, and creates a bottleneck at the root of the tree! The OpenFlow reference controller ( controller) implements a bridge/learning switch, as does ovs-controller and/or test-controller (currently the default controllers for Mininet), as does NOX's pyswitch module, and they don't implement a spanning-tree protocol by default. As a result, they will not work with a network that has loops in it. In general, if you want to use a network with loops in it, you need to be absolutely sure that your controller supports such a network. As mentioned above, ovs-controller, controller and pyswitch do not by default.

POX includes a spanning tree module, and other controllers (Floodlight, ONOS, ODL, etc.) may support multipath and/or spanning tree - you will want to consult the documentation for your controller, make sure it is configured correctly to support multipath or spanning tree, and test it to make sure that it actually works. A simple test is to use RemoteController pointed at your controller and use the torus topology, e.g.: sudo mn -topo torus,3,3 -controller remote,ip=,port= Please feel free to fill in this chart with test results from various controllers: controller version topo result details 1.0 -torus 8,8 success need to create proactive routes or start reactive forwarding Beryllium -torus 3,3 success support for looping topologies If you just want to get your network 'working', you can run STP. In Mininet 2.2 you can use the Linux bridge: sudo mn -topo torus,3,3 -switch lxbr,stp=1 In the current master branch, you can also use OVS in bridging mode: sudo mn -topo torus,3,3 -switch ovs,failMode=standalone,stp=1 or the more compact: sudo mn -topo torus,3,3 -switch ovsbr,stp=1 You will need to wait for STP to converge. You can observe its progress with sh brctl show s1 ( LinuxBridge) or sh ovs-ofctl show s1) ( OVSBridge.) You can also call net.waitConnected to wait for STP to converge: mininet py net.waitConnected Note that if you are running a remote controller (rather than a local OVS or Linux bridge as suggested here), waitConnected will only wait for the switches to connect to your controller.

If you are using a remote controller, you should check the controller console or logs for any updates. As noted above, running spanning tree removes any performance improvement from multipath networks, although it can still provide redundancy for reliability (if you deactivate a link, STP can compute a new spanning tree that uses a different link and restores connectivity.) If this sounds terrible, it's because it is - one of the advantages of using a multipath-capable OpenFlow controller is that you can potentially escape the tyranny of Spanning Tree! If you wish to code your own multipath-capable controller in POX, you may also wish to take a look at, which provides starter code for a multipath-capable controller, as well as some of the multipath experiments on. But, you will still probably have to do some work and actually understand what you are doing. Why do the switch data ports have random MAC addresses? How do I assign MAC and IP addresses to the switch data ports?

Mn Topo For Mac Free

The MAC addresses reported by Linux for the switch data ports are meaningless. The switch is controlled by OpenFlow, so you should use OpenFlow to ensure that any packets destined 'for the switch' are properly routed. You 'assign' MAC and IP addresses 'to the switch' by using OpenFlow rather than the Linux IP stack. If you attempt to use the Linux IP stack instead, it really won't work unless you are using the Linux kernel for routing (which you aren't - you're using an OpenFlow switch!) You should never attempt to use ifconfig or ip addr or other utilities to assign an IP address to a switch data port that is connected to a host or another switch ( veth interfaces are not bridges!) Usually you will want your controller to handle packets such as ARP and ICMP which are sent to and from 'the switch,' and you will want IP packets which are sent to to be handled by appropriate flow table entries. You can pick any 'MAC' address you like for the switch. Why does my network fail if I use more than 16 switches? For Mininet 2.0.0, the default controller for the mn command is ovs-controller (which can be installed automatically in Ubuntu.) Unfortunately ovs-controller only supports up to 16 switches.

If you want to use more than 16 switches, you should use a controller that supports more than 16 switches, for example: sudo mn -controller ref -topo linear,20 -test pingall Make sure you've installed the Mininet version of the OpenFlow reference controller, which is easily done using: mininet/util/install.sh -f You can also create a custom controller class or use -controller external:IP and use any custom or off-the-shelf controller that you like. For example, ou can easily install POX by checking it out or using util/install.sh -p, and you can install Floodlight on Ubuntu using apt-get install floodlight. If you are using the default controller or any controller which implements an Ethernet bridge (aka learning switch), be sure that your network does not have loops in it or that you have. How can I control my Mininet hosts remotely? It's trivial to control Mininet hosts from the CLI or from within a Python script running locally, but what if you want some other process or even another computer on your LAN to be able to control your Mininet network remotely? Well, there are lots of ways to do this. One idea is that anything you can do in Python, you can do in Mininet, and it's often very easy to do so.

For example, there are all sorts of frameworks available for any kind of messaging you can imagine. (See below for a REST example - it's just a few lines of code.) Another easy way to control Mininet hosts in the current master branch is to use the util/m script. For example if my Mininet server is ubuntu1, I can run ifconfig on host h1 using $ ssh ubuntu1 mininet/util/m h1 ifconfig Another way is to actually connect your Mininet network to your LAN and to run sshd on your Mininet hosts. This is left as an exercise for the reader, but you may want to look at the hwintf.py and sshd.py scripts in examples/ to understand how you might possibly do this. How can I add a REST interface to Mininet? Basically anything you can do in Python you can do in Mininet, and it's often very easy to do so. For example, there are all sorts of frameworks available for various kinds of messaging and RPC, REST, JSON, SOAP, XML, etc.

You can hook Python code up to node.js, you can have it speak ZeroMQ, you can use Apache Thrift. Really the possibilities are endless!!

Note however that if you are running locally it's much easier to control Mininet directly from within a Python script or using the CLI. It's trivial to add a REST (or ReST if you prefer) API to Mininet using Python. For example, using the framework, you could do something like. #!/usr/bin/python from mininet.net import Mininet from mininet.topo import SingleSwitchTopo from bottle import route, run, template net = Mininet( topo =SingleSwitchTopo( 2 ) ) @route( '/cmd// ') def cmd( node = 'h1 ', cmd = 'hostname ' ): out, err, code = net.get( node ).pexec( cmd ) return out + err @route( '/stop ') def stop: net.stop net.start run( host = 'localhost ', port = 8080 ) This allows you to send simple commands to your Mininet hosts. Note that host.pexec (like host.cmd) runs commands as root, so this isn't really something you want to expose to the whole internet (though you could firewall port 8080 on your Mininet server and then use ssh for a secure connection.) But it's quite convenient, isn't it? In a real example you would probably want a method to shut down both the REST server and the Mininet network in a graceful manner rather than using control-C and mn -c.

After running this script in one window: $ sudo./rest.py You can easily try it out. For example, you could run the ifconfig h1-eth0 command on host h1 as follows: $ curl localhost:8080/cmd/h1/ifconfig%20h1-eth0 h1-eth0 Link encap:Ethernet HWaddr 36:6f:c0:28:a3:f9 inet addr:10.0.0.1 Bcast:10.255.255.255 Mask:255.0.0.0 UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:4 errors:0 dropped:0 overruns:0 frame:0 TX packets:3 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:328 (328.0 B) TX bytes:238 (238.0 B) Note that you need to escape the space in ifconfig h1-eth0 as%20 as you would normally in a URL. How do I run the Mininet examples?.

Fetch the Mininet source tree if you haven't already git clone. Run an example sudo mininet/examples/treeping64.py Note some examples ( consoles.py, miniedit.py) require you to have set up. Others require that you have installed the OpenFlow reference implementation using mininet/util/install.sh -f.

Consult the comments in the example's.py file for details. Why doesn't dpctl work? How can I dump a switch's flow table? Dpctl should work fine with the Stanford OpenFlow reference implementation or the CPqD version of same. But you probably don't want to use dpctl at all - use ovs-ofctl instead! Especially if you're running install - ovs-ofctl it's particularly easy to use with OVS and, importantly, will actually dump the complete flow table (unlike ovs-dpctl!) If you're running Open vSwitch, or need to open up a listening port on either OVS or the reference switch so that you can connect to a port, read on. If you are using Open vSwitch, the correct command to use from the shell prompt is $ sudo ovs-ofctl dump-flows s1 (If you try to use ovs-dpctl with Open vSwitch, you will only see the cached flows in the kernel, rather than the switch's full flow table.).

The correct command to use from the Mininet prompt is mininet dpctl dump-flows which dumps all the flows on all switches, and works for both OVS and the reference switch. If you're running OVS, you can pass the switch name to ovs-ofctl and it will connect to it via the file system: mininet sh ovs-ofctl dump-flows s1 or, from the shell prompt: $ ovs-ofctl dump-flows s1. If you want to open up a listening port on the switch, you need to specify the base listening port, e.g. Net = Mininet( topo=topo, listenPort=6634 ) Ports will be allocated sequentially starting with the value you specify.

Note if you want to dump the flows from the reference switch, you will need to have a listening port opened up; then you can use dpctl: $ dpctl dump-flows tcp:localhost:6634 Note that ovs-ofctl doesn't like localhost, so you should use 127.0.0.1: $ ovs-ofctl dump-flows tcp:127.0.0.1:6634 How do I generate traffic on my Mininet network? Asking this question usually means you haven't read or understood the or indeed this. Mininet runs pretty much any Linux program. So, you can use pretty much any client or server program you can think of (e.g. Ping, iperf, wget, curl, netperf, netcat etc.) You can easily capture traffic using programs like tcpdump and wireshark. You probably should do a Google search on something like. It's also easy to generate packets in Python using scapy.

If you want to generate or decode OpenFlow messages, you should look at various controller frameworks like POX or OpenFlow messaging libraries like OpenFlowJ or LOXI. How do I modify packet headers using Mininet? Asking this question usually means you haven't read or understood the documentation and that you don't understand what OpenFlow is.

Use OpenFlow. Please go through the OpenFlow tutorial and consult the OpenFlow specification. How do I implement a custom routing algorithm? Asking this question usually means that you haven't read or understood the documentation and that you don't understand what OpenFlow is. Use OpenFlow. Please go through the OpenFlow tutorial and consult the OpenFlow specification. How do I update to a new version of Mininet?

I am setting a test topology in Mininet as the following: sudo mn -mac -topo linear,3 -switch ovsk -controller=remote,ip=10.0.2.15,port=6653 like this: I want the traffic going from H1 to H3 to go visit H2 before being forwarded to H3. I have set flow rules on my switches (Open vSwitch) to do that. However, when traffic goes to H2, it gets dropped there.

What configurations I should add to H2 (Ubuntu 14.04) in order to enable H2 to forward a traffic that is not destined to it, nor originated from it? The feature I think you'll need to enable is IP Forwarding. This can be accessed with the sysctl utility: sysctl -a grep forwarding produces (on my Mac): net.inet.ip.forwarding: 0 net.inet6.ip6.forwarding: 0 It may look slightly different on Ubuntu. The 0 value means IP Forwarding is disabled. For me to enable IP Forwarding (v4), I would, as root: nevin-mac-mini: root# sysctl net.inet.ip.forwarding=1 net.inet.ip.forwarding: 0 - 1 This should take effect immediately, however, it will not persist after a reboot. To make it so, I would place net.inet.ip.forwarding=1 in /etc/sysctl.conf.

Thank you for you answers, I finally could do it by rewriting the destination Mac address of the packets in the switch. The problem was that, the host 2 drops packets that are not destined to it ( different Mac destination address). Thus by rewriting the destination Mac address to host2's Mac address in switch 2, and by activating the IP forwarding in host 2, traffic is successfully redirected. So the step one, is to tell S2 to send traffic to Host 2 with destination Mac of host 2. And step 2, is to tell S2 to rewrite the the Mac address of traffic returning from host 2 to host 3 Mac address ( to restore the original mac address).