I've argued for some time that large-scale meshes based on Wi-Fi technology (MAC or PHY or both, and hey, let's extend this to any wireless metro-scale mesh technology just for argument's sake) would be more than successful - they'd become, in fact essential. Granted, this sounded pretty stupid a couple of years ago when the first wave of muni meshes, those on the left side of the Chasm (Geoffrey Moore's observation remains absolutely brilliant lo these many years after the book was published), largely crashed and burned. My attribution for this state of affairs at the time rested largely with untested business models rather than anything being fundamentally wrong with market demand or the technology itself. Such happens all the time, after all, and I knew mesh would be back stronger than ever.
Need evidence, here in 2011? Wi-Fi leader Aruba Networks this week announced a new line of ruggedized, outdoor mesh products, called AirMesh, at least partially based on their recent acquisition of mesh technology vendor Azalea Networks. Let's first reconsider that business angle once again: would Aruba, now a big public company, get into this space if the opportunity wasn't really high and the opportunity cost really low? Of course not; their shareholders would hit the roof and heads would roll. So they must have made a great internal business case for this announcement, and I personally don't think such was all that hard. Industrials. Verticals. Cellular offload. Emerging markets. And, yes, even municipal-scale deployments with business models based on reality and not the political goals of local bureaucrats (free Wi-Fi for all? Universal service? Governments owning these networks? Wow, did we ever really believe that nonsense? Sadly, we did.). The required technology today is far more advanced, costs are lower, and the need is greater than ever.
So Aruba has jumped in with single-, dual- and even quad-radio 802.11n-based mesh nodes, along with the required planning and management software. As is also the case with a number of other mesh vendors, vertical apps are being stressed by the firm, along with video surveillance, already a huge application of mesh all by itself. Aruba is also emphasizing its Layer-3 architecture, and I do expect some debate on this point with the proponents of Layer-2-based solutions. The fundamental problem of analyzing the performance of meshes, identical to that for wireless LANs, remains: real-world benchmarking is very difficult due to the cost involved and the fundamental variability of test environments, lab tests may not reflect real-world performance, and reliable comparative analytical models don't exist.
No matter - there's no good alternative to meshes of this type in many applications, simply based on the economic constraints involved. And more no matter - Aruba's new emphasis on mesh is a powerful indication of the bright future ahead for the technology. Mesh dead? Hardly.