# Computer Network 7｜Implement MiniNet The following instructions is based on MacOS with Apple silicon chip (M1/M2). The used VM software is VMware and Virtual Code Studio is required. ### 1. VM Setup - Register on VMware - Download: VMware Fusion Player 13 – Personal Use (VMware Fusion 13.0.1 (for Intel-based and Apple silicon Macs) https://customerconnect.vmware.com/en/evalcenter?p=fusion-player-personal-13 - Select personal use for a free license - Download from one of the following links for the latest VM image - [Dropbox Link](https://www.dropbox.com/s/tv894lhg851pzxv/cs6250-f22-vmware-arm.vmwarevm.zip?dl=0) - [Onedrive Link](https://gtvault-my.sharepoint.com/personal/jrandow3_gatech_edu/_layouts/15/onedrive.aspx?id=%2Fpersonal%2Fjrandow3%5Fgatech%5Fedu%2FDocuments%2FVM%2DImages%2Fcs6250%2Df22%2Dvmware%2Darm%2Evmwarevm%2Ezip&parent=%2Fpersonal%2Fjrandow3%5Fgatech%5Fedu%2FDocuments%2FVM%2DImages&ga=1) - Run image with the VMware installed ### 2. Basic Test - Login to VM with the following information: - username: mininet - password: mininet - Then open a QTerminal (terminal emulator) window for testing ```bash $ sudo su - # for root permission [sudo] password for mininet: mininet ``` - Finally, test mininet installed ```bash % mn *** Creating network ... exit ``` ### 3. IDE Setup - Open VSCode with Remote-SSH extension - Go to VM terminal and check for the ip address. Check the IP address (`inet`) under `ens160`. For me, it is `192.168.80.129`. ```bash $ ifconfig ens160: ... inet ... ... ``` - Connect to the host -> add a new host, then enter the following command ``` ssh mininet@192.168.80.129 ``` - Use the default config and connect. Enter the password `mininet` when required. - Wait to be connected. Then we can use VSCode to edit. ### 4. Mininet Introduction - Github Source: https://github.com/mininet/mininet - Offical webpage: https://mininet.org/ - Beginner Guide: http://mininet.org/walkthrough/#interact-with-hosts-and-switches Here are some common commands, - `help` – shows all commands - `nodes` – displays all nodes including the controller c0 - `net` – displays all nodes and links in an extended format - `links` – brief display of links - `dump` – displays full network information - `$mn –c` cleans the Mininet environment after topology runs. ### 5. Useful Sources - Network topology visualizer for `complextopo.py` and `datacenter.py` (will discuss later): http://demo.spear.narmox.com/app/?apiurl=demo#!/mininet ### 6. Project Setup - Make `Assignment` directory ```bash $ mkdir /home/mininet/Assignments ``` - Then use VSCode to go into this folder and copy the starter code from unzipping `SimulatingNetworks.zip` downloaded from Canvas - Go to the folder ```bash $ cd /home/mininet/Assignments/SimulatingNetworks ``` - Time to start! Note: you can also setup git/github for your working directory before you start. ### 7. Part 2 - Modifications If you can successfully execute the program `topology.sh`, feel free to skip the following discussions. Because I downloaded the VM image through the first link and it is not perfect, we have to make some change when we are running Part 2. Inside the `SimulatingNetworks` folder, change the permission by, ```bash $ chmod -R 755 ./topology.sh ``` Also install `bwn-ng` by, ```bash $ apt-get install bwm-ng ``` Then bypass the `cgroups` issue by adding the following line in `/etc/fstab`, ``` cgroup /sys/fs/cgroup cgroup defaults,blkio,net_cls,freezer,devices,cpuacct,cpu,cpuset,memory,clone_children 0 0 ``` Also, add the following line in `/etc/default/grub`, ``` GRUB_CMDLINE_LINUX_DEFAULT="quiet cgroup_enable=memory,namespace systemd.unified_cgroup_hierarchy=0" ``` Then update `grub` with, ```bash $ update-grub ``` And reboot the VM, ```bash $ reboot ``` Once back, rerun `./topology.sh` then we are able to execute. However, after a couple seconds, we will receive a `matplotlib` error. Then we add the following line at the top of the `util/helper.py` file ```bash #! sh/bin/env python ``` And we should also install `matplotlib` by, ```bash $ pip3 install matplotlib ``` After this, if we run `./topology.sh` again, it will pass without any further errors. ### 8. Part 2 After running `./topology.sh`, we generate the following two figures in the output folder, - cwnd: shows a TCP cubic pattern - bandwidth: shows a rate of about 10 Mbps ![](https://i.imgur.com/GKZjWnu.png) The topology here we used is defined in `mntopo.py` with 2 hosts, 2 links, and 1 switch. ![](https://i.imgur.com/qWrxGyf.png) **Task 1**: Modify the `mntopo.py` so that it represent the following topology. Rerun `./topology.sh` to get the result. Remember to clean the mininet files by `mn -c`. ![](https://i.imgur.com/RkdykBa.png) After execution, we have a similar because our topology map doesn't have much difference. ![](https://i.imgur.com/MZu70DT.png) Next, let's test the latency of the current topo. First let's run `python ./ping.py` for the initial bench mark as, ```bash $ python ./ping.py PING 10.0.0.1 (10.0.0.1) 56(84) bytes of data. 64 bytes from 10.0.0.1: icmp_seq=1 ttl=64 time=19.4 ms 64 bytes from 10.0.0.1: icmp_seq=2 ttl=64 time=9.55 ms 64 bytes from 10.0.0.1: icmp_seq=3 ttl=64 time=8.70 ms 64 bytes from 10.0.0.1: icmp_seq=4 ttl=64 time=8.85 ms 64 bytes from 10.0.0.1: icmp_seq=5 ttl=64 time=8.82 ms --- 10.0.0.1 ping statistics --- 5 packets transmitted, 5 received, 0% packet loss, time 4006ms rtt min/avg/max/mdev = 8.697/11.063/19.389/4.173 ms ``` **Task 2**: Modify `linkConfig.delay` in `mntopo.py` to `10ms`, then rerun `python ./ping.py` to compare the result. The possible result is, ```bash $ python ./ping.py PING 10.0.0.1 (10.0.0.1) 56(84) bytes of data. 64 bytes from 10.0.0.1: icmp_seq=1 ttl=64 time=166 ms 64 bytes from 10.0.0.1: icmp_seq=2 ttl=64 time=85.5 ms 64 bytes from 10.0.0.1: icmp_seq=3 ttl=64 time=81.7 ms 64 bytes from 10.0.0.1: icmp_seq=4 ttl=64 time=81.8 ms 64 bytes from 10.0.0.1: icmp_seq=5 ttl=64 time=81.5 ms --- 10.0.0.1 ping statistics --- 5 packets transmitted, 5 received, 0% packet loss, time 4007ms rtt min/avg/max/mdev = 81.537/99.397/166.426/33.546 ms ``` Here we say see pings just a bit over 80 ms compared with 10 ms initially. **Task 3**: Adjust bandwidth `linkConfig.bw` to `50` Mbps per second. Rerun `./topology.sh` to get the result. Remember to clean the mininet files by `mn -c`. The result should be, ![](https://i.imgur.com/kYBwybE.png) **Task 4**: Create a new topology representing a more complicated network topology in `complextopo.py`. ![](https://i.imgur.com/LoSELH1.png) The parameters should be set as follows, - Ethernet: Bandwidth 25 Mbps, Delay 2 ms, and loss rate 0% - WiFi: Bandwidth 10 Mbps, Delay 6 ms, and loss rate 3% - 3G: Bandwidth 3 Mbps, Delay 10 ms, and loss rate 8% Also, do not specify ports (use the default ones) in this task. ### 9. Part 3 Now that we have implemented `complextopo.py` so it's the right time to test it out. We run `cli.py` to start the simulation. ``` $ python ./cli.py ``` After Mininet loads the complex topology, you should see the Mininet command prompt: `mininet>`. Then we can test the following command, ``` mininet> h1 ping h2 -c 10 ``` This means to let `h1` to ping `h2` for 10 times. If everything works well, we can expect something like, ``` PING 10.0.0.2 (10.0.0.2) 56(84) bytes of data. 64 bytes from 10.0.0.2: icmp_seq=1 ttl=64 time=1087 ms 64 bytes from 10.0.0.2: icmp_seq=2 ttl=64 time=57.8 ms 64 bytes from 10.0.0.2: icmp_seq=3 ttl=64 time=27.7 ms 64 bytes from 10.0.0.2: icmp_seq=4 ttl=64 time=26.9 ms 64 bytes from 10.0.0.2: icmp_seq=5 ttl=64 time=25.1 ms 64 bytes from 10.0.0.2: icmp_seq=6 ttl=64 time=25.8 ms 64 bytes from 10.0.0.2: icmp_seq=7 ttl=64 time=26.2 ms 64 bytes from 10.0.0.2: icmp_seq=8 ttl=64 time=25.2 ms 64 bytes from 10.0.0.2: icmp_seq=9 ttl=64 time=25.1 ms 64 bytes from 10.0.0.2: icmp_seq=10 ttl=64 time=25.4 ms --- 10.0.0.2 ping statistics --- 10 packets transmitted, 10 received, 0% packet loss, time 9039ms rtt min/avg/max/mdev = 25.064/135.192/1086.846/317.361 ms, pipe 2 ``` **Task 1**: Test ping from `h1` to `h2` and `h3` for 100 times each. Compare the difference of the results. - h1 to h2: delay 20 ms with 4% packet loss - h1 to h3: delay 30 ms with 10% packet loss This follows the topology we have built. **Tasl 2**: Use `mininet> pingall` to ping between each pair of hosts. If success, we should expect something like, ``` mininet> pingall ... h1 -> h2 h3 h2 -> h1 h3 h3 -> h1 h2 *** Results: 0% dropped (6/6 received) ``` Note that there can be `X` in the result meaning the one link is not pingable. If this happens, please rerun `pingall` to confirm. ### 10. Part 4 In this part, we are going to implement a fan-in style data center. It has three layers of switch, - `tls`: top-level switch - `mls`: mid-level switches - `rs`: rack switches, these connct to the hosts The network should have the following two arguments, - `fi`: fin-in rate, wich is the ratio of, - the # of mls / the # of tls - the # of rs / the # of mls - `n`: number of hosts connected to each rack switch For example, ![](https://i.imgur.com/Uyhi8lw.png) The code should be implemented based on `datacenter.py` using loops. And the naming conventions should follow, - tls: tls1 - mls: mls1, mls2, to mlsfi - rs: s1x1, s1x2... to sfixfi - hosts: h1x1x1, h1x1x2, to hfixfixn For testing, we can run, ```bash $ python datacenter.py --fi 2 --n 5 ``` Then run `pingall` when we are in the Mininet prompt. We can expect this result if everything works well. ![](https://i.imgur.com/GyuJIaa.png) ### 11. Submission To submit, use the following command to zip the file. ```bash $ zip -r _sn.zip . -x '.*' -x '__MACOSX' ```