Last night, I set out on another expedition to test WiFi connectivity using an OM1P Open Mesh router and a Ubiquiti Nanostation LOCO M2. There were two main objectives. The first was to test alien node connectivity, where an Open Mesh router from one mesh network connects to a nearby mesh router from another network. The other objective was to test connectivity with the Nanostation from a distance of roughly 270 meters.
One thing I expected from mesh routers was that they would look for a link to the Internet, even if on a separate SSID, and use that if the home network is not available. I'd asked if this could be done, and got a negative; so, I considered the matter dead. It turns out that an Open Mesh router will connect to another Open Mesh router with an open network if within range, even though it has a different SSID.
In other words, if JoesWiFi cannot connect to the JoesWiFi gateway, it will jump to JudysWiFi and use that as a gateway. This is all provided that JudysWiFi also uses an Open Mesh router with open security.
To me, the ability to use alien nodes is a positive. This means that if you are building a community wireless network, neighbors can customize their settings locally; but, they can still participate in the public mesh. In this way, each network can choose a different splash page, if desired; but, neighbors can still mesh with the best available connections.
In any case, I meshed my router with a nearby gateway with a different SSID. Then, I used my mesh router to run a cable over to the Ubiquiti radio, serving as an access point.
Connecting With Uneven Gain
The second part of the experiment involved pointing the Ubiquiti radio to a park about 270 meters away. This is nothing exciting given the power of the Nanostation. However, the unit was about 4 feet off the ground and roughly aimed, hardly ideal. Even so, I was able to get 2 to 3 Mbps throughput. Again, this is hardly exciting as Nanostations can easily connect with another high gain unit over a handful of kilometers.
What makes the situation different is that I was using my Android Tablet at that distance, pulling the throughput mentioned. Obviously, a mobile device has much lower gain and power available to make the connection. To me this means that the Ubiquiti Nanostation did most of the heavy lifting in terms of sensitivity and power. Using just the mesh routers alone, the connection would have been spotty.
Given that I am footing the bill on these experiments, my next move is to try the same link using two Nanostations, one on each end. The goal is to see how high a throughput is possible from the same distance of 270 meters. After that, I'll see how far I can stretch the link from a 20 foot height in an urban setting. I have to foot the bill on each permutation; this could take a while before I can make a city-wide network as I intend to do.
Just to give you a preview of my thinking, I think the mesh routers will work more or less OK; but, there is also a need to establish backhauls to the community network. My thinking is that backhauls can add some geographic diversity to the Internet sources.
Let's say that I have a mesh router hooked up to a wireless bridge to another neighborhood. If my neighborhood's broadband connection goes down, the link could pull traffic from a different neighborhood. The same link using mesh alone would involve too many hops and network overhead. The backhauls would serve to leapfrog distant gateways.
Another idea I have is to replace the dipole antennas on the mesh radios and replace them with Yagi antennas or some other high-gain replacements to establish distant links that serve as meshed backhauls without the need for bridges. This would preserve the mesh while reducing the number of extraneous connections.
I have some crazy ideas on interconnecting stuff. The only thing holding me back is lack of capital to carry them out. But, I have time and patience. I'll keep you posted on my findings.