Meru Brings Visualization to Wireless

My last real job before I started Farpoint Group was with a supercomputer manufacturer, Stellar Computer (RIP). There I spent a lot of time working on scientific visualization, the art of mapping large, multivariate data sets onto appropriate graphical representations in hopes of quickly finding insight in what would otherwise be a tangled mess of data. The result of our work back then is still available, in an evolved form, from Advanced Visual Systems, and other firms have similar software tools.

Visualization has some great applications in wireless, and to date have mostly taken the form of simple management graphics to those interesting "heat maps" of simulated coverage produced by many wireless-LAN management tools. The idea is to graphically show where a given signal will propagate, and how it fades (usually mapped to color) assuming placement of a given AP with given transmit power and specified building components (walls, doors, building materials etc.). It can be a lot of work to set this up - importing a .dxf of the building by floor, placing APs, setting a wide variety of properties, trying multiple cases - but the results look, in most cases, anyway, spectacular.

The problem with this approach is that it's very difficult to get everything right, and thus the results, however pretty, are called into question. Is there really a 5.5 dB loss through that wall? Better hire an engineer with a good spectrum analyzer to find out. And, of course, such results say nothing about effective performance as they do not model data loads or duty cycles. Nonetheless, a good first-order approximation can often be obtained, providing some guidance as to how many APs are needed and where they should be placed.

But my approach, as many of you know, is not to do this kind of planning or even a site survey (in most cases of open-office buildings, anyway). Rather, I like to get an idea of user loads by location, data loads and applications demands, and duty cycles, all of which can be obtained from network management logs, do an initial WLAN installation based on this data, measure the result, and then add a few more APs to compensate for errors. This converts expensive, inaccurate up-front labor into cheap hardware, which the client is eventually going to need anyway as their network grows.

The post-installation measurements are again usually taken from management logs, but I sometimes will walk around with a spectrum analyzer to see just how well signals are in fact propagating. This is also expensive and mostly inaccurate, but again a good first-order approximation. Wouldn't it be great, though, to build this RF/PHY-level measurement and analysis into the WLAN, eliminating the walking, and adding a degree of accuracy that is otherwise unobtainable?

Meru Networks has now done just that with a new "virtual reality" visualization program that shows 3D representations of RF coverage based on real (not simulated) data. I've not used this product yet, but I am intrigued by the approach and I think, based on my background with visualization, that this concept is going to be a huge success. Along with RF spectrum management tools for interference management, RF environment visualization tools add another important capability that network managers will love, and should even have a positive impact on operational expense. I can't wait to try this one.

• Wireless / Mobile
• Meru Networks
• RF Environment Visualization
• RF Spectrum Management