Dili Village Telco Part 8 – April Trip Summary

Lessons Learned:

  1. Technical tasks we consider basic are difficult in the developing world. For example 2 hours to reflash a Nanostation 2 here compared to a few minutes for some one who is experienced. A short video would really help with complex tasks as it conveys subtle timing information and gets around the need for English comprehension.
  2. Special care and attention must be given to training. It must be possible for the students to repeat the exercises without any external support (e.g. via the Internet or local expertise) once the teacher leaves.
  3. Basic configuration of the Mesh Potato (e.g. setting the IP/Phone number) is easy to teach and works well. It meets the ease of use vision of the Village Telco. However setting up reliable mesh links needs to be dramatically simplified. There are many complex variables that currently require a high level of skill.
  4. A way to determine if long power cables are causing problems would be very useful. This problem is compounded by the need for high masts. Power problems can be very difficult to diagnose or debug.
  5. Power being switched off or going down is a common problem. For example some one would simply unplug the MP power supply and a link would go down. Unlike a cell network, we depend on end users to keep their Mesh Potatoes and hence the network up.

    One subtle issue are socket adaptors, for example to convert between US and local style power sockets. These are often of poor quality and cause intermittent problems.

    Servers need to tolerate going up and down several times a day, and tolerate issues like a DHCP server or Internet link being unavailable without freezing.

  6. Keeping the mesh up for 24 hours can be a major challenge. People unplug power, masts move, or operators just don’t check if the network is running.
  7. Don’t rely on the Internet. It can be very slow or unavailable. Don’t rely on contact with your server over the mesh. Your system should work fine without the Internet and provide fall back capability (for example local node-node calls) without contact with a central server.
  8. Interference can be a big problem in urban areas of the developing world. Poor DSL infrastructure means everyone is using Wifi to distribute Internet or bridge LANs across town. To fix interference problems requires a high level of skill and expensive, hard to source equipment. More work is required to make interference problems easier to diagnose and correct.
  9. There is a trade off in mesh Wifi between omni antennas and interference. You need omni for the mesh, but omni antennas have poor performance in areas of high interference. Interference problems can be solved with directional antennas or routers with directional antennas, however this equipment is expensive and hard to procure in the developing world. David Carman has suggested the idea of a MP antenna that can be omni or directional via a slide in reflector. This may help from a cost and procurement perspective.
  10. Mesh node installation can be a major engineering exercise and potentially dangerous due to the heights involved and local safety standards. For typical links of a few hundred metres line of sight is required. If you have 15m trees, this means 15m masts. In flat terrain it’s not as simple as the “simple pole a few metres above the roof” that I had visualised. This has been a problem for me in two mesh networks I have deployed on flat terrain. However networks deployed in hilly terrain have some natural height that can be put to good use. In low interference environments, non line of sight paths are more viable as low signal strengths can be better tolerated.
  11. Local calls (MP-MP) are the killer application for the people in Timor Leste that I spoke to. There was little interest in gatewaying to the GSM network or overseas calls via VOIP. In areas with GSM coverage many GSM calls are to the same number, or even to the office down the hall. GSM call costs are so crippling people will sacrifice the convenience of a mobile phone. In areas with no GSM coverage the ability to make a simple local phone call can save hours of travel along poor roads, and potentially save lives.

Links

Dili Village Telco Project Wiki
Dili Village Telco Part 1
Dili Village Telco Part 2
Dili Village Telco Part 3
Dili Village Telco Part 4
Dili Village Telco Part 5
Dili Village Telco Part 6
Dili Village Telco Part 7

2 thoughts on “Dili Village Telco Part 8 – April Trip Summary”

  1. Inspiring and super interesting to read of your experiences over there. I definitely wouldn’t have expected some of your lessons learned.

    Great stuff.
    Simon.

Comments are closed.