Pyxnet: SDN testing library

Authors:	Florian Dupeyron <[email protected]>
Date:	December 2022

Introduction

Pyxnet is a python (py) library targetted at creating virtual networks (net) using SDN [1] technology. The x in the name is for the crossover thing. Its main application is for testing software behaviour in a given environment, or external devices like embedded devices. This tool is structured in such a way that it is simple to integrate in automated test environments, and that created network topologies are easy to export in documentation.

Pyxnet goals

[1]	Software Defined Networking

Installation

You can install this package via pip:

pip install git+https://github.com/fdmysterious/pyxnet

Dependencies

This library depends on the following elements:

• Linux only
• openvswitch must be installed, and the following commands must be available:
  • ovs-vsctl
  • ovs-dpctl
  • ovs-ofctl
• iproute interface is made using the pyroute2 library.
• graphviz allows graph generation

Defining a network topology

General principles

A network topology is defined using the following elements:

• Topology objects
• Links between objects' endpoints

For instance, Let's take the following topology:

Example topology (examples/basic_topology.py)

Toplogy objects are the four switches s1, s2, s3 and s4, each switch have two endpoints p0 and p1, and links are symbolized by the lines on the graph.

The python code generating this topology is fairly straightforward:

if __name__ == "__main__":
  logging.basicConfig(level=logging.INFO)

  # First, declare a topology object. This object will holds all objects declaration,
  # as well as links between these objects.
  tt = Topology(name="Basic topology")

  # Secondly, we declare the objects we will have in our topology.
  s1 = tt.register(MyCustomSwitch("s1", mac_addr="02:01:02:00:00:01", ip_addr="10.0.0.1/24"), group="group1")
  s2 = tt.register(MyCustomSwitch("s2", mac_addr="02:01:02:00:00:02", ip_addr="10.0.0.2/24"), group="group1")
  s3 = tt.register(MyCustomSwitch("s3", mac_addr="02:01:02:00:00:03", ip_addr="10.0.0.3/24"), group="group2")
  s4 = tt.register(MyCustomSwitch("s4", mac_addr="02:01:02:00:00:04", ip_addr="10.0.0.4/24"), group="group2")


  # Then, we connect these objects together
  tt.connect(s1.p0, s3.p0)
  tt.connect(s1.p1, s4.p0)
  tt.connect(s2.p0, s3.p1)
  tt.connect(s2.p1, s4.p1)

  # Then, we can export our graph
  dot = tt.export_graphviz()
  dot.render("output_graph")

  # Now, onto instanciation of our topology!
  cleanup_all()    # Precaution
  tt.instanciate() # Topology instanciation on host platform

  # Up objects
  s1.up()
  s2.up()
  s3.up()
  s4.up()

This example is available in the examples/basic_topology.py file. We can see here that after the topology object is created, the topology is defined by registering objects into our topology, then linking them together. Objects can be grouped in logical groups for easier representation.

Endpoint types

Each toplogy object defines endpoints, which are the points that can be linked together in order to define how objects are linked together. This corresponds for instance to a switch's port. Each endpoint can have the following type:

• Virtual: the endpoint exists in the virtual world, e.g. inside the host machine;
• Real: the endpoint exists in the real world, e.g. outside the host machine;
• Phy: the endpoint is an interface on the host machine, e.g. a usb/phy dongle, or a port on an ethernet card.

This type definition imples the following statements:

• All the network topology can be defined, virtual and real objects;
• A virtual endpoint cannot be linked directly to a real endpoint. There must be a phy interface in-between.

Real and virtual objects (examples/real_objects.py)

As defined, the user doesn't have to worry about implementation details, and what object neeeds to be defined or not, wether it have to be instanciated in the linux machine or not. This also makes the generated graph exhaustive about the tested topology.

On the above figure, we can see that links existing in the virtual world are solid lines, whereas links in the real world are represented using dashed lines. This means that when instanciating, real links have no real effect on what's happening inside the linux host.

Defining a custom object

Defining a custom object is straightforward, as it is class based. Here is an example of how to define a custom switch with two ports:

class MyCustomSwitch(Switch):
    """
    This class illustrates how a custom network object can be created
    by simply inheriting base objects
    """

    def __init__(self, name: str, mac_addr: str = None, ip_addr: str = None):
        # Each topology object has a name
        super().__init__(name,
            mac_addr   = mac_addr, # Not mandatory
            ip_addr    = ip_addr,  # Not mandatory

            stp_config = {
                "rstp_enabled": True,     # Enable RSTP!
                "bridge_priority": 0x8000 # Set bridge priority
            }
        )

        # Init endpoints
        self.p0 = self._endpoint_register("p0", Endpoint_Kind.Virtual)
        self.p1 = self._endpoint_register("p1", Endpoint_Kind.Virtual)

        # Set endpoint RSTP properties
        self.p0.properties["stp_config"] = {
            "path_cost": 100,
            "priority":  0x8000,
        }

        self.p1.properties["stp_config"] = {
            "path_cost": 100,
            "priority": 0x8000,
        }


    def export_graphviz(self, dot):
        """
        Illustrate how it's possible to customize
        the diagram generated object
        """

        dghelpers.box_logo_node(dot, self.name, dghelpers.asset("icons/material/router.png"), f"Switch {self.name}")

What we can see here, is that the important elements are:

• Each network object has an unique name ;
• the export_graphviz() method exports the object on a graphviz diagram. The google material icon set is included in this package.

To define a network object from scratch, you can derive the PyxNetObject class. The following methods can be defined:

• instanciate(self): Implements how the object is instanciated on the linux platform;
• remove(self): Implements how the object is removed from the linux platform;
• up(self): Implements how the object is bring up on the the linux platform;
• down(self): Implements how the object is bring down on the linux platform;
• export_graphviz(self, dot): Implements how the object is represented on a graphviz diagram.

License

MIT License

Copyright (c) 2023 Florian Dupeyron <[email protected]>

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.