To run the Distrinet client, you can clone this repository and run Vagrant.

git clone https://github.com/jdoe79250/algo_experiments.git
cd algo_experiments
vagrant up

after few minutes, the vm should be running, you can connect to it with:

vagrant ssh

The environment will not configure a physical infrastructure for you. To install the environment you can follow the official Distrinet website:

https://distrinet-emu.github.io

https://distrinet-emu.github.io/general_environment.html

Following the tutorial, you can run the example using the default RoundRobin Mapper with:

ryu-manager /usr/lib/python3/dist-packages/ryu/app/simple_switch_13.py --verbose

as explained in the tutorial

for the new algorithm implemented, you have to create the file description of the infrastructure. The vagrant VM create an example environment.

The example file describe an infrastructure composed by 2 physical machines with IP_1=192.168.0.10 ip_2=192.168.0.11, connected with one switch sw1.

The example file is located at:

/home/vagrant/.distrinet/simple_environment.json

You can adapt this file for your infrastructure.

For Network experiment you have to enable the PREBUILD = [default_images] command in the file "/vagrant/algo_experiments/experiments_fattree.py"

This command will run the network experiment in FatTree topology, K=4 and links at 500mbps, all the vNodes requires 1 vCore and 2Gb of RAM

bin/dmn --controller lxcremote,ip=192.168.0.1 --bastion="192.168.0.10" --workers="192.168.0.10,192.168.0.11"  --custom=/vagrant/algo_experiments/fattree/fattree_topo.py,/vagrant/algo_experiments/experiments_fattree.py  --test piperf --topo ft,4,500,1,2,1,2

For CPU and Memory experiment you have to enable the PREBUILD = [default_images, toHadoop] command in the file "/vagrant/algo_experiments/experiments_fattree.py"

This command will run the network experiment in FatTree topology, K=4 and links at 500mbps, all the vHosts requires 8 vCore and 12 Gb of RAM, while the vSwitch requires 1vCpu and 2 GB of RAM

bin/dmn --controller lxcremote,ip=192.168.0.1 --bastion="192.168.0.10" --workers="192.168.0.10,192.168.0.11"  --custom=/vagrant/algo_experiments/fattree/fattree_topo.py,/vagrant/algo_experiments/experiments_fattree.py  --test hadoop --topo ft,4,500,8,12,1,2

This command will run the network experiment in FatTree topology, K=4 and links at 500mbps, all the vHosts requires 8 vCore and 12 Gb of RAM, while the vSwitch requires 1vCpu and 2 GB of RAM

Be careful on the ramdisk and file size in the hadoop_mem() function

bin/dmn --controller lxcremote,ip=192.168.0.1 --bastion="192.168.0.10" --workers="192.168.0.10,192.168.0.11"  --custom=/vagrant/algo_experiments/fattree/fattree_topo.py,/vagrant/algo_experiments/experiments_fattree.py  --test memory --topo ft,4,500,8,12,1,2

You can chose the mapper, addin the --mapper parameter to the command.

for example for embedded greedy you can run:

bin/dmn --controller lxcremote,ip=192.168.0.1 --bastion="192.168.0.10" --workers="192.168.0.10,192.168.0.11"  --custom=/vagrant/algo_experiments/fattree/fattree_topo.py,/vagrant/algo_experiments/experiments_fattree.py  --test piperf --topo ft,4,500,1,2,1,2 --mapper=embeddedGreedy

You can choose between this mappers:

• roundRobin
• maxinet
• embeddedBalanced
• embeddedGreedy
• embeddedPartition
• random
• switchBin