Virtual Network Generator (vngen)

Automated virtual network generation in golang utilising QEMU virtulisation and KVM hardware acceleration

• Command line based application
• Automated network virtualisation with KVM and QEMU
• YAML-defined network-host topology configuration
• Create, run and manage multiple deployment topologies
• Data persistence
• RestApi server mode for remote connection.
• Log management
• Supported on Linux only

Table of Contents

• Requirements
  • Ubuntu/Debian
  • Arch Linux
  • Ubuntu/Debian
• YAML Topology Configuration
• Command Line Interface
  • Installation
  • Create Network Deployment
  • Start, Stop, Restart and Destroy Hosts or Deployments
  • Display Information
    • Hosts
    • Networks
    • IPs
• Rest API Server
  • Server (Localhost mode with http)
  • URL Endpoints
    • Build
    • Start, Stop, Restart, Destroy
    • Details

Requirements

Install KVM and Qemu

Ubuntu/Debian

sudo apt install -v qemu libvirt-client libvirt-dev cloud-utils

Arch Linux

sudo pacman -S qemu libvirt cloud-utils

Ubuntu/Debian

YAML Topology Configuration

Create custom topologies with yaml templates. Simply add networks, hosts and then assign networks to hosts as per the default.yaml example:

---
# Deploymnet Configuration
deployment: 
  name: default

# Network Configuration
networks:
  - name: br0
    netaddr: "20.0.0.1"
    dhcplower: "20.0.0.2"
    dhcpupper: "20.0.0.254"
    netmask: "255.255.255.0"
    type: "nat"

#Hosts Configuration
hosts:
  - name: master1
    image: ubuntu
    ram: 2048
    cpus: 2
    hd: "10G"
    username: dev
    password: ved
    networks:
      - br0
  - name: master2
    image: ubuntu
    ram: 2048
    cpus: 2
    hd: "10G"
    username: dev
    password: ved
    networks:
      - br0

Command Line Interface

Installation

# Clone repository
$ git clone https://github.com/Kenjasim/vngen.git
$ cd vngen 
# Build application binaries
$ go build -o vngen cmd/vngen/vngen.go 
# Add vngen application to PATH
$ mv vngen /usr/local/bin

Create Network Deployment

sudo vngen build </path/to/template> # default.yaml

Start, Stop, Restart and Destroy Hosts or Deployments

sudo vngen start [deployment|host] <name>
sudo vngen stop [deployment|host] <name>
sudo vngen restart [deployment|host] <name>
sudo vngen destroy [deployment|host] <name>

Display Information

You can display hosts, networks, and IPs

sudo vngen get hosts
sudo vngen get networks
sudo vngen get ips

Hosts

[i] Getting hosts
Name    VMState Image  RAM  CPU Storage Deployment 
master1 off     ubuntu 2048 2   10G     default
master2 off     ubuntu 2048 2   10G     default

Networks

[i] Getting networks
Name Type IP       DHCP Range            Deployment 
br0  nat  20.0.0.1 20.0.0.2 - 20.0.0.254 default

IPs

[i] Getting ips
Name    Interface MacAddr           IPs        Deployment 
master1 vnet0     52:54:00:2b:76:8f 20.0.0.141 default
master2 vnet1     52:54:00:ed:90:f9 20.0.0.88  default

Rest API Server

The RestAPI Server enables remote access to the application either through AVN's client mode, or via direct http (localhost), https (remote) requests.

Server (Localhost mode with http)

To launch the server just run the command on the binary

sudo vngen api

# Or you can specify the port

sudo go run cmd/vngen/vngen.go api -p 30675 

URL Endpoints

Build

You need to send a PUT request with a jsonified version of a YAML template file, an example can be seen below:

{
    "deployment": {
      "name": "default"
    },
    "networks": [
      {
        "name": "br0",
        "netaddr": "20.0.0.1",
        "dhcplower": "20.0.0.2",
        "dhcpupper": "20.0.0.254",
        "netmask": "255.255.255.0",
        "type": "nat"
      }
    ],
    "hosts": [
      {
        "name": "master1",
        "image": "ubuntu",
        "ram": 2048,
        "cpus": 2,
        "hd": "10G",
        "username": "dev",
        "password": "ved",
        "networks": [
          "br0"
        ]
      },
      {
        "name": "master2",
        "image": "ubuntu",
        "ram": 2048,
        "cpus": 2,
        "hd": "10G",
        "username": "dev",
        "password": "ved",
        "networks": [
          "br0"
        ]
      }
    ]
  }

You can then pass this to the endpoint as a post request with the header set as application/json

http://localhost:8000/build

Start, Stop, Restart, Destroy

These 4 commands follow the same trend, all must be sent as POST requests and all have a simmilar structure.

http://localhost:8000/<start|stop|restart|destroy>/<deployment|host>/[name]
http://localhost:8000/start/deployment/default

Details

To get a list of all defined hosts or networks you can use this URL endpoint:

http://localhost:8000/hosts
http://localhost:8000/networks

To get more details about one host you can use:

http://localhost:8000/details/[name]

Finally, to get the IP of a host you can use:

http://localhost:8000/details/[name]/ipv4

These should all be run as GET requests