Making metro-scale Wi-Fi meshes fly

Multiple radios solve myriad Wi-Fi mesh problems

Metro-scale Wi-Fi networks are one of the most important trends in wireless, so not surprisingly, I've written about it quite often lately. However, Ive received so many questions via e-mail that a bit more clarification is in order.

I believe that metro-scale Wi-Fi networks, based almost exclusively on meshes, will become ubiquitous over the next few years -- and on a global basis. Imagine mobile, broadband communications operating (mostly) in common frequencies and deployed virtually everywhere using unlicensed bands and mass-marketed radios, which will keep provider and consumer costs low. This is real 4G wireless, well ahead of the cellular guys!

One frequent complaint about this vision is, in fact, those unlicensed bands. However, as I noted in a previous column on this topic, interference shouldn't be a show-stopper here. Meshes are designed for resilience, working around interference and other threats to service in a manner transparent to users. Besides, hundreds of megahertz are available around the globe, and not every channel will see interference all the time. Good spectrum management techniques and the higher transmit power used in most infrastructure meshes should handle that problem quite nicely in most cases.

But the question remains: Can a mesh really provision enough capacity for multiple classes of users and multiple classes of applications, all on a single infrastructure? Some mesh equipment vendors are tackling this challenge by building mesh nodes with multiple radios; its not unusual to see six or even more radios per node.

Multiple radios are important because each additional radio means both greater reliability and more capacity. More capacity means more users can be on the air simultaneously, more data can be moved per user, or both. And since the demand for mesh capacity will be no different than the demand for network capacity in general, which means it will only grow over time, the ability to economically expand a mesh is critical. This means that, for example, field-upgradeable mesh nodes have a real advantage in holding both capital and operating costs down.

The other big cost element is in backhaul. The real beauty of a mesh, from an architectural perspective, is its ability to cover a large area quickly and economically because each node doesnt require a wired interconnect. Internode communications are carried over the mesh like any other traffic.

The problem occurs when we need to bridge outside of the mesh via wired or wireless (but not Wi-Fi, in this case) backhaul connections. These can be expensive, and having a small number of radios per node demands more backhaul. That's because it is costly, both in terms of capacity and latency, to relay through multiple nodes to reach a backhaul connection point. Having more radios per node eliminates this problem, allowing fewer (although higher-performance) backhaul links to be provisioned.

This is a bit complex, so Ive written a white paper that you can find here explaining all of this (and a lot more) in significant detail. Regardless, though, have no fear: Metro-scale meshes, properly architected, designed and implemented, will be a part of your wireless solution set for both personal and business applications far into the future.

Copyright © 2007 IDG Communications, Inc.

7 inconvenient truths about the hybrid work trend
 
Shop Tech Products at Amazon