zahrasf / dragon Goto Github PK

Updated Jul 30, 2018

DRAGON is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version. DRAGON is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with DRAGON. If not, see http://www.gnu.org/licenses/.

This repository provides the prototype of an architecture that cenables coexistence of different applications over the same shared edge infrastructure. To dynamically partition resources to applications, this project uses the Distributed Resource AssiGnment OrchestratioN (DRAGON), an approximation algorithm that we designed to seek optimal resource partitioning between applications and guarantees both a bound on convergence time and an optimal (1-1/e)-approximation with respect to the Pareto optimal resource assignment.

The repository tree contains:

• README.md
  --this file
• LICENSE
  --GPLv3 license
• main.py
  --main instance executable
• config/
  --configuration files
• dragon_agent/
  --DRAGON agent source files
• resource_assignment/
  --implementation of the applications-resources assignment problem
• scripts/
  --useful scripts related to the project
• tests/
  --validation purpose scripts
• use_cases_simulations/
  --simulation environment that runs two edge use case over DRAGON

Download the source code by cloning the repository, as well as any submodule:

$ git clone https://github.com/netgroup-polito/dragon
$ cd dragon
$ git submodule update --init

config/default_config.ini -- main configuration file
config/config.py -- agent and instance configuration
config/rap_instance.json -- resource assignment problem instance values

Please create a copy of the main configuration file, and edit your configuration according with your own setup:

$ cp config/default_config.ini config/config.ini

This project requires python 3.6 and has been tested on Linux debian (testing) with kernel 4.16.0-2-amd64.

Some additional python packages are required:

$ sudo apt install python3-pip
$ sudo pip3 install -r requirements.txt

Some tests require the metis library, please install it from source:

$ pip3 install Cython
$ cd [dragon]/tests/utils/nxmetis
# python3 setup.py install	

Inter agent communication is implemented over the RabbitMQ Broker. To install it use the following command:

# apt install rabbitmq-server

Make sure rabbitmq is running:

# service rabbitmq-server start

The main.py script runs a single instance of the DRAGON agent. To run it, use the following command from the project root directory:

$ python3 main.py {agent-name} {services} [-d {configuration-file}]

where:

• agent-name: is a name to identify the agent;
• services: is a list of parameters, namely the names of services for which the agent will attempt to obtain resources (see config/rap_instance.json).
• configuration-file: is the path of the configuration file to use (default is config/default-config.ini).

The tests/ folder also contains a script that automatically runs multiple agents at the same time. Please modify config/default-config.ini as desired before to run it, so to specify instance parameters, then use:

$ python3 -m tests.test_script

The number of agent specified in the configuration file (each with a random number of services) will be run and the script will wait for convergence. At the end of the execution, the log file of each agent will be available in the main folder, while details on the resulting assignments will be stored on the (generated) results folder.

An alternative script allows you to perform tests while running agents on remote hosts. Since this requires to setup ssh connections with the remote hosts, please install the ssh server on each of them:

# apt install openssh-server

then please setup your ssh public key to be accepted on every target host.

You may need to increase the limit of ssh connections accepted on each host, by modifying the 'MaxStartups' parameter in the sshd configuration file:

# nano /etc/ssh/sshd_config

Then, assuming you want to allow up to 50 connections, change the 'MaxStartups' line as follows:

MaxStartups 50:30:100

Close and save the file, then restart the ssh daemon:

# service sshd restart

Make sure rabbitmq is running on every host thorugh federated setup (see https://www.rabbitmq.com/federation.html).

The tests/test_script_ssh.py script can be setup specifying the list of remote hosts, the username to be used for the ssh connections and the remote path where dragon is located. Please modify these values in the first lines of the script according to your setup.

Analogously to the local test script, you can run the remote test using:

$ python3 -m tests.test_script_ssh

The script will automatically copy the local configuration to the remote hosts, and dragon agents will be spanned on them (agents that are neighbors in the topology file will be likely deployed on the same - or on a near - host). All output and log files will be fetched and results displayed locally.