There are some fantastic tools out there for detailed design of fibre networks from companies like Comsof and IT Simplicity but they are sometimes overkill when all you want to gauge is if fibre is even the right technology to consider or if a project is worth going to the extra level of detail. However, there are some excellent Open source GIS tools out there with more features than most of us will ever need to consider – but are there enough to get that draft fibre network design to suggest whether fibre is the right solution?
The quick answer is “yes” but not in one place.
The first really useful resources is the downloadable Open Streetmap (OSM) data and PostGIS 2.0’s pgrouting tools.
By loading the OSM data into a Postgres table and formatting it for routing allows you to create a set of roads and paths to focus on for your project.
You will, of course also need a table containing the premise locations. In the UK, at least, there is no fast and free way of doing this so its a case of either buying a copy of Addressbase from the Ordnance Survey or manually creating a PostGIS layer in QGis, adding pins for each premise using Google Maps satellite as a backdrop.
Buffering this road network allows you to create a possible trench layout with a possible trench on either side of the road.
Breaking the trench network into regular lengths and using a “nearest neighbour” script allows you to connect each premise to the nearest trench point; at this stage it makes sense to include a test to make sure the distance from each premise to the nearest trench point is not too far and is affordable.
This creates a maximum possible trench network with all the drops but also all the roads, footpaths, and bridleways – there are still a lot of trenches which don’t pass premises or which would take a longer route to each premise. This is the point to introduce QGIS and its interfaces to GRASS.
Load the trench and premise data into QGIS, and use it to create new Grass Mapset, importing the two layers so Grass can see them. Then use the v.net family of modules to find the minimum trench network, removing all the superfluous routes.
I’ve found v.net.steiner produces a reliable result, and once its finished you can quickly see which premises are connected to the network and how long the trench network is likely to be it giving you an initial view on whether a fibre to the home network is feasible.
It won’t be a perfect design and it will have chosen a few odd paths to get to some premises but it will give you a reasonable idea of what will be involved – how many kilometres of digging will be required.
This sounds like its a cumbersome process but the PostGIS tasks can be done in a single SQL script, and the QGIS work is just two steps. Once set-up it doesn’t take long to produce a quick first view of a possible solution.
This won’t be good enough to formally cost a project and it makes no attempt at a Bill of Materials or contractor work sheets but it does help you see whether a project is possible and which premises are likely to affordable.