Farpoint Group –
In my last column, I talked about micromeshes and how to build a solution out of them using gateways. This week, as promised, we'll talk a little more about meshes in metro areas, particularly those using Wi-Fi. Since my last discussion of meshes, I've received a lot of questions that indicate that I needed to expand a bit on this topic. While I think that MIMO is currently the most important radio technology (in terms of its influence on products going forward), I think meshes are the most important radio architecture. Meshes solve so many problems, but one question that I've received over and over again regarding meshes concerns their economics. While the elimination of the requirement for interconnect (usually called backhaul) between mesh nodes will clearly have a very positive effect on the cost of a mesh-based solution of any form, the examples in our new white paper on meshes show how really dramatic the cost savings can be.
But nothing is ever perfect in the world of wireless. There's been a good deal of controversy lately about the potential performance impacts of real-world loads on mesh-based wireless networks. The problem is simple - most mesh nodes today consist of only one radio, meaning that the radio can be either communicating with a client or relaying traffic on behalf of a client (and only on one radio channel), but not both at the same time. We call this store and forward. With large numbers of nodes and large numbers of transactions (or other high throughput requirements), the overall capacity of the mesh and the latency involved in moving data from point A to point B through the mesh become major issues.
Fortunately, there is a solution, especially when we use Wi-Fi-based meshes. We can, in fact, use multiple radio channels simultaneously - there are, for example, 27 channels currently defined and allowed in the US. The problem has been in putting multiple radios into a single box - a phenomenon known as adjacent-channel interference comes into play here. In this case, the energy from a channel right next to another channel (in the spectrum) causes mutual interference between the two. It's particularly a problem when the radios are very close to one another, as might be the case if we attempted to put two or more radios into a single box or otherwise operate them in close proximity.
So, if we're going to use multiple radios simultaneously in a single mesh node, we've got to be cleaver - and products with said cleverness are now becoming available. For example, a WLAN chip company called Engim produces a chipset that implements three Wi-Fi radios in one - and provides the necessary isolation and filtering so that adjacent-channel interference is eliminated. So, imagine putting an Engim chip in a mesh node, and using the three channels available for some combination of user traffic and backhaul. Available bandwidth goes up, latency goes down, capacity increases, and everyone goes home happy. Oh, sure, such a solution is a little more expensive, but such is always the case - we expect to pay for more capacity, and "more" in this case really is a very small sum indeed.
Anyway, take a look at the white paper and let me know what you think - you're going to be hearing a lot more about meshes over the next few years. Next week, though, we'll take a break from meshes as we explore RFID in a little more detail.