Rural WiFi Mesh Networks, part 2

As I mentioned, I think that a mesh network should form the “backbone” of your farm or ranch network. I have to admit I found the entire subject somewhat intimidating when I came to it, so I want to present the alternatives to you to highlight the tradeoffs, so you can make the best choice for your operation.

There are two general types of mesh networks: single-radio (store-and-forward) and multi-radio (continuous). The difference is that a single-radio system receives a “packet” of data, stores it, then turns the radio to transmit, and transmits that packet. As a result, each node on a store-and-forward system reduces the effective bandwidth by half (since it can only be receiving half the time or transmitting half the time). This means that the speed of the network diminishes the further away from the “Gateway” node (the node attached to your router) you get, and that the expansion of the network is limited. If the bandwidth between the nodes is fast enough and the range of the nodes is far enough, however, this may not be a significant limitation.

A multi-radio system has one radio dedicated to “receiving” data and another radio dedicated to “transmitting” data. By cutting out the “store and forward” aspect, both radios can run at full speed all the time, so there is no loss of bandwidth. This provides the advantage of ensuring maximum bandwidth throughout the network and enabling the network to be arbitrarily expanded. However, since the “receiving” and “transmitting” radios have to be on different frequencies, it requires a LOT more planning and expertise to set up this kind of a network. And, of course, since there are two radios and some intervening electronics, each node costs a lot more and takes a lot more power. In fact, many of these systems will have a third radio as a dedicated access point (so that bandwidth is not coming out of one of the “mesh” radios) and even a fourth radio for redundancy on the mesh or the access point.

The multi-radio mesh was originally invented for military and commercial use, where large numbers of “clients” are using large amounts of data and failure can be disastrous. These mesh nodes, besides being very complex to set up and install, are usually very expensive (thousands of dollars per node). Cisco, Motorola, Aruba, Meshdynamics, Ruckus, and several others market these systems. Meraki and Anaptyx also have multi-radio mesh networks (Anaptyx uses the Open-Mesh platform).

Some years back, the folks at MIT decided they needed a campus-wide WiFi network, and they invented something they called Roofnet. Many of the folks who created it went on to found Meraki, which is still the leader in low-cost mesh networks. Meraki, as noted earlier, has also branched out into higher-cost mesh networks. Meraki started out with an “open” platform for experimentation, and attracted hackers who wanted to develop mesh networks for their own purposes. When Meraki “closed” their software, a group of coders developed Robin in open source, which became the basis for Open-Mesh and for AyrMesh.

On a personal note, when we were starting Ayrstone, I evaluated both Meraki and Open-Mesh. Meraki, at the time, was “locked” into “client isolation,” so it did not create a Local-Area Network. This was a deal-breaker, although the higher cost of the Meraki nodes, the closed architecture (meaning we couldn’t do anything to make the nodes behave better) and the requirement of a subscription payment for their management software also served to dissuade me. Open-Mesh overcame those objections, but was only available on small, low-power, indoor-only access points – not suitable for our vision of a rural mesh network. The Meraki and Open-Mesh “dashboards” are also very full-featured, which is to say, too complex for a networking “beginner” to understand and use confidently.

So we undertook the task of building our own firmware (using ideas from Robin and another open-source mesh firmware package called “Nightwing“) to run on the very good, high-power PicoStation access points from Ubiquiti. We also developed our own greatly simplified “dashboard” at ayrmesh.com so that “normal people” could install and manage a “household” (or “farmhold” or “ranchhold”) wifi mesh network.

The bottom line is that any of these solutions (and there are even more out there) will work to build a wireless LAN on your rural property. Which one to choose depends on what you want to do. For absolute simplicity on a typical farm, ranch or rural home, I think AyrMesh is the best choice (but, of course, I’m biased). Open-Mesh and Meraki are also perfectly viable choices (Meraki even used to offer a solar-powered outdoor node, although I no longer find it on their store), especially for those who have more ambitious goals than just having WiFi on the homestead. Cisco and Motorola have long been considered the “gold standard” for this technology, and Aruba and MeshDynamics are kind of the “fast followers,” but they’ll require a certified installer and tens of thousands of dollars. Somewhere between these options lies Ruckus, which is fast developing an outstanding reputation among the Silicon Valley wireless networking folks. If I were looking to build an “enterprise-class” WiFi network, I’d certainly include them in the evaluation.

For now, I think that’s enough about the technology… next, I want to dive into implementation – “how to do it” for the do-it-yourselfer.

Leave a Reply

Your email address will not be published.