Millimeter Wireless Last Mile

MB 600 and 1200 SanFrancicso California (2015-054)I saw a webcast recently given by Siklu, an Israeli firm that makes what they call millimeter wave radios, which in the US are the V Band (57 – 64 GHz) and the E Band (71 – 66 GHz and 81 – 86 GHz). The company claims that their radios can be used to bring gigabit speeds to rural areas. Apparently a number of other people saw the same presentation because I received a number of emails and phone calls asking me if this could be a residential last mile solution.

The short answer is that their radios can do what they claim, but that this is not yet a solution for serving end-user customers in rural America. There are a number of issues that make this a poor solution today.

First, these are point-to-point radios. That means you need to hang a transmitter dish on a tower and another at each customer for this to work. Contrast this with point-to-multipoint radios where one transmitter on a tower can communicate with a few hundred customers. Anybody that has ever tried to lease space on towers knows that tower owners won’t allow a huge number of transmitters and would charge a fortune even if they did. Even for a carrier that owns towers, the idea of putting hundreds of dishes on one tower presents all sorts of logistical problems.

Then there is the issue of rain fade. These frequencies more or less die during heavy rain. Siklu claims this the average place in the US gets about 3 days per year of rain. But that’s not 3 continuous days, but the sum total over a year of an hour here and an hour there. I know here in Florida that we get evening rain storms almost daily during the summer and when we can’t go outside we turn to online entertainment.  A technology that dies when it rains would be hated here (which is how Floridians feel about satellite TV for the same reason).

The technology would fare a lot better in places where it doesn’t rain much. The only way to compensate for the rain fade is to send the signals for relatively short distances, less than half a mile, and that really reduces the usefulness in rural settings. And distances that short means spending a lot of money on fiber and on neighborhood towers.

Because of the nature of the frequencies used this is also a pure point-to-point technology. That means there can be nothing between the transmitter and receiver like leaves. Millimeter waves are really short and just about everything disrupts them.

Siklu is pushing this as a solution to bring gigabit speeds to rural customers. And it could do that if you can overcome the tower-crowding issue, the rain fade and the distance issues. But trying to bring that much bandwidth to end-users also means bringing a huge amount of bandwidth to the towers – and backhaul costs a lot more in rural areas than it does in metropolitan areas.

This finally leaves the issue of cost. Millimeter range radios from different manufacturers seem to cost $3,000 – $4,000 per pair. That sounds like a lot, but might still be cheaper than building rural fiber where there is no density. But when you add tower costs and other operating costs this technology is going to be expensive. And unlike fiber, which might last for 50 – 75 years, one would have to think the operational life of these radios is a decade at most, meaning periodic and costly replacement.

Siklu is not the only company making radios for the upper frequencies. Traditional companies like Fujitsu also make these radios along with a number of smaller companies like LightPointe and MI-Wave. But it looks like Siklu is the only one pushing this as a solution for rural broadband.

This is not to say that there are not uses for high-bandwidth point-to-point radios. The primary benefit of these radios is that they can transmit more bandwidth than traditional microwave radios. These might be a great alternative to fiber to bring bandwidth to a rural school or business (or to anybody willing to pay for a large commercial bandwidth product). This could be a great alternative to building fiber in really rough terrain or for getting big bandwidth from a valley floor up to a cellular tower on a hill. And Siklu has used these radios as backhaul for WiFi radio systems. But for all of these applications rain fade is still going to be an issue. At the distances that Siklu touts, the radios would be in service 99% of the time (two 9s). But carrier class products require four 9s (99.99%) or five 9s (99.999%) of up-time, and I find it unlikely that cellular companies or schools will tolerate the outages during storms.

I hear all of the time that wireless is going to be the solution to our rural bandwidth problem. But when you look at the practical issues associated with this and other wireless technologies in rural areas it’s easy to see that we are still a long way from having a big-bandwidth wireless technology that will work and be cost effective. These same upper frequencies will perform a lot better in urban environments where the distances are much shorter and companies like Google and Starry are probably going to make this frequency work in cities. But the life-cycle costs to make this work in the rural environment have to rival the cost of building fiber directly.

2 thoughts on “Millimeter Wireless Last Mile

  1. Dear Doug:
    The short answer is that Millimeter radio waves will experience the same issues — or worse — as the early LMDS tests in the early 2000’s.
    Customers and end-users must beware of rosy predictions and sunny California case studies — most of the rest of the country is NOT like California, Nevada, Arizona or Texas!
    By the way, I was at Virginia Tech a month ago for a conference, and no one there had heard of what Va. Tech. was doing with it…

  2. I heard from Siklu who wanted to point out that they have mitigated the rain fade issue to some extent with what they call adaptive modulation. This allows the radios to reliably deliver for distance greater than what would be predicted by the physics – 1/4 to 1/2 mile.

    They pointed to a recent Siklu blog where the radios are being used for wireless back-haul. That is an application that makes more sense for these radios instead of a last-mile alternative.

Leave a Reply