Skip to main content

Network traffic modeler API written in Python 3

Project description

PyPI Build Status Coverage Status Documentation Status

pyNTM: network_traffic_modeler_py3

This is a network traffic modeler written in python 3. This library allows users to define a layer 3 network topology, define a traffic matrix, and then run a simulation to determine how the traffic will traverse the topology, traverse a modified topology, and fail over. If you've used Cariden MATE or WANDL, this code solves for some of the same basic use cases those do. This package is in no way related to those, or any, commercial products. IGP and RSVP auto-bandwidth routing is supported.

In pyNTM, the Model objects house the network topology objects: layer 3 Nodes, Circuits, Shared Risk Link Groups (SRLGs), Interfaces, etc.

There are two subclasses of Model objects: the PerformanceModel object and the newer FlexModel object (introduced in version 1.6).
Starting in version 1.7, what used to be called the Model class is now the PerformanceModel Class. The former Parallel_Link_Model class is now known as the FlexModel class.
There are two main differences between the two types of objects:

  • The PerformanceModel object only allows a single Circuit between two layer 3 Nodes; while the FlexModel allows multiple Circuits between the same two Nodes.
  • The Performance Model will have better performance (measured in time to converge) than the FlexModel. This is because the FlexModel has additional checks to account for potential multiple Circuits between Nodes and other topology features.

The legacy Model and Parallel_Link_Model should still work as they have been made subclasses of the PerformanceModel and FlexModel classes, respectively.

The PerformanceModel class is good to use for the following topology criteria:

  • There is only one link (Circuit) between each layer 3 Node
  • IGP-only routing and/or RSVP LSP routing with no IGP shortcuts (traffic source and destination matches LSP source and destination)

In some cases, it's completely valid to model multiple Circuits between Nodes as a single Circuit. For example: in the case where there are multiple Circuits between Nodes but each Interface has the same metric and the use case is to model capacity between Nodes, it's often valid to combine the Circuit capacities and model as a single Circuit. In this case, the FlexModel object is recommended as it will give better performance. If it is important to keep each Circuit modeled separately because the parallel Interfaces have different metrics and/or differences in their capabilities to route RSVP, the FlexModel is the better choice.

There are two main areas where we are looking to optimize:

  • Performance - converging the model to produce a simulation, especially in a model with RSVP LSPs, is intensive. Improving the time it takes to converge the simulation results in better productivity and improved user experience. Possible Cython implementation or PyPy interpreter could add value here.
  • Data retrieval - the simulation produces an extraordinary amount of data. Currently, the model is only retaining a fraction of the data generated during the model convergence. It's our goal to introduce something like an sqlite database in the model objects to hold all this information. This will improve user experience and allow SQL queries against the model object.


Install via pip:

pip3 install pyNTM

For upgrade:

pip3 install --upgrade pyNTM


See the training modules on this repo's wiki page at


See the documentation on Read the Docs.


See the example directory.


Copyright 2019 Tim Fiola

Licensed under the Apache License, Version 2.0:

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for pyNTM, version 2.0
Filename, size File type Python version Upload date Hashes
Filename, size pyNTM-2.0-py3-none-any.whl (81.9 kB) File type Wheel Python version py3 Upload date Hashes View
Filename, size pyNTM-2.0.tar.gz (59.9 kB) File type Source Python version None Upload date Hashes View

Supported by

Pingdom Pingdom Monitoring Google Google Object Storage and Download Analytics Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN DigiCert DigiCert EV certificate StatusPage StatusPage Status page