Aruba Networks Brings Layer 3 To Wireless Mesh Networks

Wireless mesh networks have always been big on promise, but often disappointing on performance. It's no secret that the marketing folks for wireless vendors can be a bit adventurous with range and throughput claims, and that many wireless mesh products just don't live up to the hype. Part of the problem has been the traditional repeater-style "lose half your throughput with every hop" principle, which can be exacerbated by poor network design. Aruba Networks addresses pretty much all of the potential mesh wireless pain points with its recently introduced AirMesh product line, and even non-Aruba customers should take note.

While Aruba Networks is hardly a newcomer to the Wi-Fi world, the company's AirMesh product line roll-out does give a whole new dimension to Aruba's offerings. On the one hand, Aruba is just another thin-wireless solution, often butting up against Cisco and Meru at RFP time for those in the market for a controller-based WLAN. But on the other, Aruba has gone in a whole new direction with the AirMesh line. Controllers? AirMesh no habla controllers. In this space, Aruba competes with the likes of Firetide and Tropos, where access points hang on poles and look far more industrial than their stylish indoor cousins. The new wrinkle is that AirMesh doesn't play by typical mesh wireless rules, and its divergence from the norm is a big win for anyone contemplating big, high-performing outdoor wireless networks.

AirMesh is Aruba's new name for the recently acquired Azalea Networks portfolio, as explained to me by Greg Murphy, VP and general manager of Outdoor, Mesh and Industrial at Aruba. AirMesh continues Azalea's unique Layer 3-based approach to mesh wireless, and promises to deliver high-quality video as a simple matter of course. AirMesh's product line features quad-band radio offerings, software-defined radio technology, and a range of progressive features to go with the advantages of nodes that are as much routers as they are access points. By routing, the shortest path is always found, the need for packets to travel back to a switch or gateway prior to forwarding in the mesh are eliminated, and Spanning Tree-like topology problems are non-existent. Given that the entire product framework was designed to optimize video over the network while eliminating past mesh detractors, all network types gain traffic benefits from the AirMesh approach.

Murphy clued me in to not only the magic under AirMesh's hood, but also how two companion offerings keep AirMesh designs high-performing. An outdoor 3-D planning tool helps optimize network designs. More importantly, the Aruba Wireless Mesh Professionals (AWMP) certification program makes sure that those designing mesh networks for mines, ports, municipalities, and other large and challenging environments actually know what they are doing. As with indoor wireless, if you don't know what you are doing at the system design stage, you can easily make a great product set seem like junk, and the stakes are even higher in big mesh applications.

I pressed Murphy on what the long-term future for Aruba is, now that the company is essentially supporting two operating systems. While some of the more impressive features from the AirMesh OS may be ported over to the Aruba OS, the two systems will stay distinct for the foreseeable future. At the same time, Aruba's AirWave management system will soon be upgraded to manage AirMesh routers and networks as well as Aruba's controllers and APs. The fact that AirMesh is a distinct product set is important, as non-Aruba customers shopping for best-of-breed mesh wireless now have another option from a "big name" wireless vendor. With ever more wireless cameras being deployed and industrial control applications hitting the network, mesh wireless that actually delivers "real network" performance will continue to gain in importance. I'd very much expect to see routing meshes become the de facto standard over older Layer 2 approaches. Anyone can get a signal out there, but performance mesh? That's another story altogether.