What does it mean for FTTx projects?
When designing an FTTx network, there are a lot of decisions to make, and a lot of information to make those decisions about.
Information like “don’t build down this road” or “give 2 fibers to 172 Robertson Street, and 10 fibers to 1 Albert St” can all be contained in geospatial data. What this allows FTTx network planners and engineers to do is lay out all of their decisions in a geospatial context, and then apply some logic to connecting up the design area.
All of the different pieces of geospatial data in your area make up the candidate network. Every road and aerial span is a candidate path. Every street corner, pole and pit is a candidate node in your network. When you create a design, you do this by selecting the best candidate paths for your cable and the best candidate nodes for your splices and splitters.
This is an example area of baseline data that we use to create an candidate network for our algorithms to automatically generate an FTTH design.
Choosing the best nodes and paths is a tricky process. You need to connect up all of your customers in a way that is both constructible and cost effective. By designing your whole network in a geospatial context, you can effectively meet these needs.
This is an example of the FTTx design output that we build from that geospatial data. It need some work (fixing demand etc, but gives us a much more reliable base for FTTx Planning)
One of the biggest challenges we face when using GIS data, are all the different ways that people use to create and use it. Over time, we’ve become experts at handling GIS data to use in FTTX design. We still get some data that makes us shake our heads, but we know that with some clever maths and careful editing, we can prepare the candidate nodes and paths for any network.
3 musts for good geospatial data on an FTTx Project
1.0 Understand the data that you need
Data is the foundation to a successful fiber project, however, it’s important that you don’t overindulge. A lot of people can fall into the trap of delaying their first design because they’re missing some data, for example the latest pole or pit data in their network. It’s often possible to infer pole locations from strand data, or even street centrelines to obtain a good cost estimate for cable lengths, equipment and cable footprint.
Ultimately, what you should be trying to define are the goals of your fiber project, and determining the data that you need to achieve those goals.
2.0 Find the right data source
Open Source Data Options
There are more and more free data options opening up to the world. This is awesome for the likes of planning and feasibility as we can rapidly create the underlying candidate network for automation to build on.
- Openaddresses.io is a great free, editable source of address locations
- Openstreetmap is a free, editable map of the world that includes buildings
- Geocoding services like Google and property and regional data from Zillow and the U.S. Census create a rich picture of premises that could be connected.
Free Data is also getting better with smart tech. For example, a Facebook project is using artificial intelligence to convert satellite imagery into geolocated population data.
Third Party Data Options
There are a bunch of commercial data aggregators out there in the market that help provide a one stop shop for highly accurate data. Stemming from the first point though, it’s critical that you understand what data you really need or you could end up spending a lot of money for little reward. Some of these include:
First Party Data Options
The attitude to data to date has been very focused around the concept of if I source it myself, it will be perfect. The approach has always been to generally oursource to an unskilled workforce that can take the time to check each and every asset within a network. The problem with this is that it leaves a lot of things upto the individual. More people, all working in a fragmented way, results in chaotic results.
When it comes to ensuring higher quality first party data, it’s critical that the engineering and Field Validation teams are supported by tech, like Work@Scale, that highlights the importance of ensuring the right information is collected at each asset.
On the other side of things, companies such as CCLD Technologies are reducing the need for workers to go into the field and instead are focusing on drones and machines to validate network assets.
3.0 Set the right data structure
We’ve found that the best way to maintain good data structure structure is to curate the flow of the data.
At the onset of a project we want to ensure that all of our data is being aligned to proper naming conventions, that align to the actual process, and that everyone in the business can understand.
After the underlying input data schema is set, you want to do the same with the network soft, and hard rules. What is the capacity of certain Cabinets? What is the Maximum length of drop cables, What should we do around landmarks or Railroads? By creating a common understanding of these rules, not only can you ensure all engineers are following strict guidelines, but you can better support automation across your network deployments.
Finally you need to create the point of validation for any piece of work. By critically defining what is acceptable, and what isn’t you can keep stakeholders honest, but you can also better manage costly mistakes, and fix any issues in a quick and holistic way.
Ultimately, by structuring your data in a standardized way, you’re ensuring a much more transparent, and interoperable workflow across the end-to-end delivery of an FTTx network.
Using data to power FTTx design automation
Automated design is a bit like using a ghostwriter as opposed to writing an entire book yourself. With autodesign, an engineer or network planner can focus on telling the machine what they want (business and architectural rules) and then let the machine (software) lead the way in generating the design. Importantly, engineers can impart their wisdom to the machine, so that less technical users can generate designs using a one click solve.
As you can probably understand from this blog, the biggest challenge with auto-design is managing the candidate network (or underlying data). Auto-design tools treat the candidate network as a source of truth, so if there’s a mistake in the candidate network (for example, an address point in the wrong place), this will be reflected in the design. When there is good data an automated approach can significantly benefit not just the design for a given area, but it provides ancillary benefits for other partners and contractors as you have a clear understanding of your network.
If you’re keen to learn more around a data-led approach get in touch and we’ll talk it through with you.