for width check



ng test: {{ m.test }}

Eigenvalues 01:  Residential LoS


The viability of any wireless last-mile solution is critically dependent on the channel conditions it faces,
especially the availability of line of sight connections.

Residential access technology is seeing renewed attention, driven by continued bandwidth demand growth (4K, unicast, etc.), copper exhaustion, and negligible improvements in the high deployment costs of fiber.  A variety of new wireless concepts are currently being explored and developed as alternatives.

The viability of these wireless options is critically dependent on the characteristics of links feasible in real-world deployment scenarios, since this profoundly affects one of the most significant cost drivers for a wireless network:  site density.

At the request of our customers, we’ve recently analyzed a random sample of 20 residential deployment scenarios across the US and 15 in Europe to help inform conversations about these wireless options with the facts.  Our study indicates that the availability of line-of-sight (LoS) links is typically quite limited in the real world of residential neighborhoods — on the order of only 10 to 20% of  sites within a 500m radius — suggesting that much care must be taken in assessing the practical deployment economics and difficulty of using any technology that depends on LoS transmission.  [Note that Tarana solutions do not.]

We provide here a brief summary of this analysis and its results.  You may certainly contact us for more details.


We started with a random sample of 20 of the top 100 markets in the United States, resampled iteratively until reasonably distributed coverage of north, south, central, east, and west geographies was achieved.


We added another random sample of 15 of the top 45 markets in France, Germany, Italy, Spain, and the UK (3 per country).

Each of these 35 markets was then sprinkled (courtesy of Excel’s random function) with four randomly-placed map pins within its metro area in order to choose a residential neighborhood to analyze.  In most cases only one of the four pins actually landed in or near a residential or mixed neighborhood — where more than one did, we flipped a coin.  For each randomly-chosen neighborhood, a 500m reference sector and a 10/15/20m ground-based or 2m rooftop installation pole was fabricated in kml objects and “installed” at one corner of the development or neighborhood in Google Earth for visual analysis.

Note that before this analysis approach was pursued in volume, the accuracy of foliage and vertical structure feature representations in Google Earth (very critical components of the picture) was calibrated for a small number of sample neighborhoods.  While the visual representation can lack photorealism and some detail, from the perspective of overall size and location of foliage and building features, Google Earth's synthesis of multiple satellite images to create 3D representations is accurate enough to make this LoS prevalence assessment approach directionally correct.


The analysis process involved manually counting the total number of buildings in the 500m reference sector and then counting the buildings by feature class (window, eave, or rooftop) that could be seen from the Google Earth camera eye positioned at each of the installed 10, 15, 20m, or rooftop pole levels, on a large, high-resolution screen.  Analysis of these feature classes was necessary to adequately assess wireless approaches that depend on different mounting conditions.  Window, eave, and rooftop mounting requirements all present materially different degrees and types of installation difficulty, and as the analysis shows, equally materially different frequency of LoS availability.


In 7 of the 15 European neighborhoods, the prevalence of tall MDU structures made rooftop installation the most appropriate scenario in which to assess LoS reach.  The intermediate 10/15/20m pole heights were not analyzed in these cases.

All together, the study included 8,219 occupied structures in the 35 neighborhoods, and 3,736 individual observations of feature classes in LoS.




The results summarized below will be easier to understand intuitively with a closer look at a representative example.  In case after case in this analysis, the first property-counting step in the process, done using a straight-down aerial view that emphasized wide-open streets and lots of rooftops, left an impression that the neighborhood would likely yield ample LoS connections.  This view of a neighborhood in Salt Lake City, Utah is an excellent example.


Once the camera’s eye is brought down to a base-station level view for this same neighborhood, however, the picture changes completely, and the low incidence of LoS connections becomes much more intuitively obvious — driven by a combination of obstructions by buildings, foliage, and variations in terrain.


As noted at the outset, the aggregate observations indicated very limited LoS availability, especially in the most realistic deployment scenarios in the US — neighborhoods dominated by single-family dwellings, with no rooftop mounting, and modest base-station height.


The higher average building density in the sample for Europe (295 per sector v. 173 in the US) created somewhat more LoS favorable-conditions, as the higher density tended to crowd out large trees more so than in the US, but LoS links remained by far the minority of cases in the 500m sector because of the persistence of building- and terrain-based obstructions.  Note that the results here are largely consistent with European LoS probabilities empirically derived for channel modeling in ITU-R M.2135.

These results have profound implications for the deployment economics of wireless solutions that are either working in spectrum where reflection and diffraction effects suffer from heavy signal attenuation, poorly equipped to leverage reflections and diffractions without undue counterproductive effects from self- and unmanaged-interference and disturbance from motion in the channel, or both.

We’ll look at these adverse network economics effects in a subsequent Eigenvalues entry.  Meanwhile, feel free to contact us if you’re interested in learning more about the details of this study.


Rocket science inside.Unrivaled performance outside.