Software

The ANS lab is currently engaged in the realization of an emulator of wireless mesh networks, or networks in general.

The research we carried out in the netCommons project showed that a distributed network can be used to provide connectivity to hundreds, or even thousands of people, and that it grows “organically”: the network expands when new people join it, which limits the need for initial expenditure and reduces the chances of market failure. What we don’t know yet is to what extent a CN, with a unique mix of technology and bottom-up organization can represent not only a filler for market failure areas, but a radically new model for network deployment. We need new models to study and forecast the growth of bottom-up networks that can help scientists and policy makers compare their performance against other options (e.g. fiber, 5G, TVWS…) even out of market failure situations. And finally, we need new performance metrics that do not deal with ‘bit/s’ only but express the social impact a network can produce.

The emulator we are currently developing is an attempt to fill this gap. Methodologically, the emulator starts from open data (demographics, income, street map, terrain, building altitude, connectivity coverage, devices performance etc.) to produce a data-based network generator that mimics the growth of a network in a specific place, with the constraints given by a specific technology (wireless/wired, planned/unplanned etc.). The generator can be used to estimate the network cost, performance, and geographical penetration of a number of wireless and even wired technologies, taking into account not only the performance of the link to the user (which is what wireless planning tools already offer) but the performance and scalability of the whole network. The results of the generator can be then abstracted into analytical models in order to estimate network performance even in areas for which we don’t have all the layers of open data.

The long term goal in this research is to propose a new theory for network design and network planning that includes non-economic factors in the process of network engineering. We need a new framework to help answer the question: “what is the technology to connect currently disconnected people (or to upgrade their current connectivity) which guarantees the highest social impact?”.

We are at the beginning of this project, we have a running back-end, and a front-end able to simulate some of the characteristics of a real network. Our initial results are included in a paper under review in the ad hoc networks Elsevier journal and the source code is available on github.

Right now, given a couple of points we are able to assess, for regions for which we have data available, if there is line of sight and the estimated bandwidth using Ubiquiti radios. We let the network grow as a sequence of random points and we can generate topologies. To have a quick view of what can be done, check this map we generated at random, or this animation that shows a potential evolution of a mesh network.

PeerStreamer-ng is a fully-fledged application for real-time P2P streaming. Its development started in the Napa-Wine European project and continued in the netCommons project. It is made up of a P2P powerful but lightweight streaming core and a client side flexible interface. Its main features are:

• Open source, self-contained C code
• Modular, built with extensibility in mind
• Advanced bleeding-edge Peer2Peer technologies (check out the bibliography)
• Created for mesh network users (check our partner projects below)
• Support for modern browser technologies: HTML5, WebRTC

Check PS-ng website for more information, source code and related scientific papers.