dtn-tvg-util

This project enables the creation and processing of Delay- and Disruption-tolerant Networking (DTN) scenarios in the form of contact plans and time-varying network graphs. This functionality can be taken as a basis for DTN simulation tools. dtn-tvg-util supports Python 2.7 and Python 3.5+.

Requirements

• Python 2.7 or 3.5+
• a build toolchain for your platform plus the python-dev package to install the Python dependencies
• to place Ring Road ground stations, libgeos-dev and libgdal-dev

Getting Started

For just using this module, install it via pip:

pip install dtn-tvg-util

For the development setup, you should first create a virtual environment and install with pip install -e:

python3 -m venv --without-pip .venv
curl -sS https://bootstrap.pypa.io/get-pip.py | .venv/bin/python
source .venv/bin/activate
pip install -e .

Some example tools can be found in tvgutil.tools. These scripts allow to generate and convert time-varying graph representations of Ring Road [1] scenarios.

For using the Ring Road scenario generators, you have to install some further dependencies:

pip install "dtn-tvg-util[ring_road]"
# OR
pip install -e ".[ring_road]"

For ground station placement (via tvgutil/tools/create_gs_list.py), also install the gs_placement extras by adding [ring_road,gs_placement] to the pip install command.

Usage Examples

You can find an example of the scripts being applied in the aiodtnsim repository. The aiodtnsim simulation script leverages dtn-tvg-util modules for providing contact plans to the simulator and its routing algorithms.

Testing

Unit tests for the core functionality are provided in test/. They are based on the Python unittest package. Use a test runner of your choice to execute them, e.g. pytest or nose.

pip install nose
python -m nose test

Concepts

Contact Types

dtn-tvg-util differentiates within its tvgutil.contact_plan module between two types of contact plans and their associated time-varying graph (TVG) representations:

• Factual Contact Plan (FCP): A list of Contact instances, representing the actually-happening contacts (as they are observed by the nodes). This representation could be obtained, for example, by observing a real network topology and recording the concrete links. An FCP is expected to be used, e.g., by an event-based simulator to represent the actual contacts that are encountered by the nodes over the course of the simulation.
• Predicted Contact Plan (PCP): A list of PredictedContact instances, representing the contacts predicted by a node. Every node may have an individual PCP. This representation is supplied to routing algorithms such as Contact Graph Routing (CGR).

Note: In earlier versions, there was only one type of contact plan as the module was not aimed at testing different DTN routing algorithms, but only theoretical analyses of the network graph had to be performed. However, this has changed with the addition of more and more simulation functionalities.

In fully-deterministic scenarios, i.e., if all nodes always use an accurate contact plan, a PCP can be converted into an FCP and vice versa. In that case, the predicted characteristics are equal to the factual characteristics and there is only one generation of predicted characteristics.

TVG Representation

The tvgutil.tvg module adds another representation of a network scenario in the form of a "time-varying network graph". Factual as well as predicted contact plans can be converted into a TVG representation. It is the preferred serialization format for scenarios generated by this package.

Overall, both a P-TVG and an F-TVG instances contain three properties:

• vertices: A set of nodes part of the graph.
• edges: A dict mapping tuples of (transmissing node, receiving node) to a list of contacts.
• contact_type The type (i.e., Contact or PredictedContact) of the contact instances contained in the values within the edges dictionary. his specified whether the TVG is a P-TVG or an F-TVG.

Transmission Plans

Most existing DTN simulation toolchains generate messages (or bundles) "on the fly", while the simulation is being run. If randomness is involved, which is typically anticipated to add some variations to the simulated scenario, the simulation result becomes non-deterministic and is not easily reproducible (if not considering "workarounds" such as fixed random seeds).

dtn-tvg-util promotes another approach: A "transmission plan" can be generated by the tvgutil.transmission_plan module. This list of messages to be generated is later provided as input to a simulation tool, which creates the messages at the times specified in the plan. This way, every simulation run is fully deterministic and can be easily reproduced given the same input data. If variation is needed, several transmission plans can be generated for a single scenario.

License

This code is licensed under the MIT license. See LICENSE for details.

References

[1] Ring Road Networks (RRN) are Disruption-tolerant Networks (DTN) for communication, based on the deployment of low-cost LEO satellites. For more information, see the following paper: S. C. Burleigh and E. J. Birrane, “Toward a Communications Satellite Network for Humanitarian Relief,” in International Conference on Wireless Technologies for Humanitarian Relief, Amritapuri, India, 2011.