Using Config Files

Fogbed accepts a topology.yml file to build an Experiment.

To create a FogbedExperiment, define the following sections: containers, instances, and links. Furthermore, you have the option to create a single topology by setting the is_distributed section to control when the topology should be distributed across the workers.

Example of topology with .yml file:

topology.yml

is_distributed: false

containers:
  node1:
    dimage: ubuntu:trusty
    dcmd: /bin/bash
    environment:
      DATA_PATH: /tmp/data
    ports:
      - 80: 8000
    resources: medium
  node2:
    dimage: ubuntu:trusty
  node3:
    dimage: ubuntu:trusty
  node4:
    dimage: ubuntu:trusty

instances:
  cloud:
    model: 
      type: cloud
      max_cu: 32
      max_mu: 2048
    containers: ['node1']

  fog:
    model: 
      type: fog
      max_cu: 8
      max_mu: 512
    containers: ['node2']

  edge:
    model: 
      type: edge
      max_cu: 1
      max_mu: 128
    containers: ['node3', 'node4']

links:
  cloud_fog: 
    delay: 10ms
  fog_edge: 

workers:
  worker1:
    ip: hostname1
    port: 5000
    reachable: ['cloud']
    instances: ['cloud']

  worker2:
    ip: hostname2
    port: 5000
    reachable: ['fog']
    instances: ['fog', 'edge']
    links:
      fog_edge:

tunnels: ['worker1_worker2']

To run that topology use the command:

fogbed run topology.yml

Tip

If you set is_distributed: true that example will run the experiment using two workers.

Make sure the IP addresses and ports are configured according to the experiment script and run a worker in each machine with:

fogbed worker

Using the ExperimentBuilder

experiment.py

from fogbed import ExperimentBuilder

if(__name__=='__main__'):
    exp = ExperimentBuilder(filename='topology.yml').build()

    try:
        exp.start()
        node1 = exp.get_docker('node1')
        node3 = exp.get_docker('node3')
        print(node1.cmd(f'ping -c 4 {node3.ip}'))

        input('\nPress Enter to exit...')
    except Exception as ex:
        print(ex)
    finally:
        exp.stop()

To run that experiment use the command:

fogbed run experiment.py