Tales from the Towers, Chapter 45: I’ll give you two 5s for 10

In this article, Rory Conaway argues why wireless ISPs should use the 10 GHz frequency for backhaul. He urges the FCC to open it up for unlicensed use.

What if there were a new 500 MHz wide band of spectrum barely being used that wasn’t on anyone’s radar? What if that band already had equipment sold by one of the best known manufacturers in the WISP industry? What if that band could use 802.11 chipsets, meaning that you don’t have to get a second mortgage on your house to put up an access point (AP) and a few clients? What if cows could fly? Then I probably wouldn’t have had the opportunity to experience the Wendy’s Triple*, but that’s not important (Airplane is one of my favorite movies). What’s important is where this mythical land of spectrum lies, how we get access to it, and how it benefits WISPs. I’ll give you a hint: just stick a GHz at the end of the title. Or if you are a Star Trek fan, say “10GHZ, The Undiscovered Country”.

5 GHz has been the backbone of WISPs for a long time. Since 900Mhz has been under assault from every direction, including the FCC, the latter becoming almost useless for WISPs in cities and suburbs. This means you are limited to only serving Grizzly Adams neighbors and hope that most users only want to watch 1970s TV shows on NetFlix.

3.65 GHz WiMax is tough on the budget and gives you ulcers, but it’s still the best option for long-range transmission and areas with vegetation. It’s just not magic. 2.4 GHz is pretty much toast in any area, city or suburbs. If you really want to see interference, I’ll show you the logs from our Xirrus router installed indoors that picked up 1400 APs. 2.4 GHz still has value in remote areas for WISPs due to the 3-1 exception rule, but without Vivato, nobody has an AP that can take advantage of it. I’m still hoping that with 802.11ac and some down conversion to 2.4GHz, the idea makes a comeback. Even though 802.11ac can handle up to 8×8 streams, it’s probably not going to happen since it would be proprietary. If the FCC would allow the 3-1 rule to be applied to the 5 GHz bands, including the DFS bands, that would be awesome since 802.11ac could take advantage of it.

The FCC is picking up the ball finally, and might grant another 195 MHz in the band, with some limitations, to expand the use of unlicensed bands. With the only part of the band, UNI-II Upper, having sufficient EIRP to penetrate vegetation limited to 100MHz, it’s limiting the use of 5GHz as a backhaul frequency on a tower that already has 4 APs running at 20 MHz. Another 100 MHz will bring back the option of 5GHz as a backhaul frequency on the same tower without needing GPS. Coincidentally, that’s just in time for 802.11ac, meaning very high capacity unlicensed backhaul (see Chapter 43: Tales from the Towers – Galactus, Destroyer of Wireless Worlds, okay, a cheap plug).

Many tower deployments that were originally installed with 5.8 GHz backhaul had to change over to licensed backhaul since the 5.8 GHz band was needed for the client connection. As Dolly Parton said at a rather embarrassing public wardrobe failure, you can’t put 10lbs of potatoes in a 5lb sack (wise woman, that Dolly). Since DFS frequencies simply don’t have the power output of a Nextel cell phone (You have to like a cell phone that can make CRTs within 10 feet, scramble the screen when a call comes in. Now imagine what it’s doing to your brain two inches away), UNI-II upper is the only band that makes tower deployments profitable in rural areas. With 5.8 GHz clients 10 miles away or more, that means towers that could be 20 miles apart. Unfortunately, 24 GHz, the next unlicensed band over 5.8 GHz, won’t be going that far unless it’s mounted on the moon.

Licensed frequencies, where most WISPs have to go to, are not the best option but they are the only option. If you are a small WISP, Part 101 rules mean you won’t be getting a licensed link for several weeks and you will be spending at least $10K or more. Although several distributors and manufacturers have made the process as painless as possible, there is still bureaucracy involved which starts at about $1500 and goes up to about $3000, depending on the deployment. Then there are the radios themselves which start at about $5K. I understand why these frequencies need to be regulated and licensed, but for small, growing WISPs, it’s a difficult pill to swallow.

Using the 10 GHz band for backhaul

So, what do WISPS do for additional cost-effective backhaul in the US? How about 10 GHz? I’m a big fan of inexpensive equipment that can be built from 802.11 chipsets and offers almost the range of 5 GHz and this would be the perfect combination. Ubiquiti sells a product like this for about $700 per radio/antenna. Although that’s a little expensive compared to 5 GHz 802.11 backhaul radios, it’s still one of the cheapest radios around in that band. Unfortunately, we can’t use this radio in the United States yet.

Internationally, the 10.0 to 10.5 GHz band has been in use with great success. Gateway Communications, a Nigerian WISP, uses it for their national frequencies. Much of Europe also uses it for point-to-point (PTP) backhaul links. In the United States, the International Amateur Radio Union uses it for ground and satellite communications, but in reality, there is very little use in those bands. That means up to 500MHz of bandwidth could be used for WISPs and other data applications.

The problem is nobody ever asked the FCC to open it up to the unlicensed band community in the United States. Brian Hinman, CEO of Mimosa Networks, saw an opportunity and has formally applied to the FCC to open up the band. If FCC approves the application, we will all benefit from having another band in which to expand our backhaul options. Unfortunately, that decision lies only with the FCC. Since the WISP community went 0-5 in the last game, this would be a chance for the FCC to change the impression that they are something other a division of the incumbent cellular and wireline providers (That ought to endear me to them further since my last few articles had such a glowing review of their performance. I call them like I see them and what I see really stinks of bureaucratic incompetence and influence. Now they will probably revoke my CB radio license and send the IRS over to audit my Wendy’s receipts).

10 GHz isn’t a perfect band by any stretch. It’s almost twice as bad as 5.8GHz when it comes getting through vegetation. The costs of the radios are also going to be higher initially, not because of a huge cost of manufacturing, but because there will be fewer of them sold in the beginning. Simply adding a 1 Gbps port instead of a 100 Mbps on a PoE port is at least $10 at the manufacturing level and $30-$40 by the time it gets to those of us who think it should only add $5 to the retail price. Adding another frequency multiplier circuit is also going to be more expensive. Ubiquiti’s 10 GHz radio, the PowerBridge M10, can handle 50-60 Mbps TCP/IP average and the cost is about $350 per radio.

But for those of you who forget that the price of a product has to include the cost of engineering and that manufacturing in small quantities isn’t cheap, that’s a reasonable price. If you toss in what the competitors charge, and let’s face it, manufacturers will charge the most they can for a product because they are in business to make as much profit as possible, then there is no reason or ability to bring that costs down unless the market expands. Other players in that band like Proxim and SAF, could cost $6K-$15K.
If the FCC opens up part of the 10 GHz frequency band to unlicensed use, where does that leave the WISP community? It depends on the limitations placed by the FCC regulators on an industry that works pretty well without a lot of government intervention. If UNI-II upper rules are used, with or without DFS capability, then the band becomes critically important. The first major advantage to filing as outlined by Hinman is that the band means the use of 802.11 chipsets. That avoids all the software-defined radios and custom chips that have driven the cost of White Space radios so high; the ROI is longer than the motorcycle I bought to save money on gas. At least that’s the argument I used with my wife when I explained that the Yamaha R1 is a very, very fuel-efficient vehicle because you barely have to twist the throttle to reach highway speeds. For some reason her acceptance to the motorcycle idea was about as good as the WISP industry is to White Space so far.

If the FCC puts UNI-II lower specifications, it becomes almost useless. The real value on this band is a low-cost backhaul alternative to licensed frequency. To be that, it needs to be able to reach at least 10 miles or more with 256QAM capability. If it can go farther than that, it will be a huge boon to the industry as well as to municipal governments and private industry. With the amount of bandwidth available in the band and 802.11ac supporting up to 1 Gbps or more, this is a game changer. 802.11 based equipment simply isn’t going to cost anywhere near what it would cost for a 1 Gbps licensed link at those distances. If fact, with Ubiquiti’s AirFiber setting the standard in 24 GHz at $3K for 700 Mbps and SAF, Trango, and others announcing similar products at $5000 or less, an 802.11 based 10 GHz at a true 400 Mbps or greater should be cheaper and create another level of product line as disruptive as AirMax and AirFiber. Even if it’s the same price, $3K-$5K, it can move AirFiber speeds at far greater distances. The fastest 802.11N products right now (from Radwin, Cambium, and others) are still limited to about 200 Mbps and prices range from $3k-$20K. This means prices for the 10 GHz versions of these products be about the same or more expensive.

Hinman is also proposing additional features to address some of the shortcomings of the UNI-II upper band. In addition to DFS which is needed to avoid some of the systems that are already in the band, they are also suggesting a contention-based protocol for spectrum reuse and taking it out of the hands of the operators. Canopy WISPs currently use a combination of frequency allocation and GPS synchronization but that’s voluntary and doesn’t work between different vendors. Unfortunately, it’s not in the 802.11 spec so even though it’s good in the long run, it might slow down equipment availability since vendors would have to go back and figure out how to deploy it.

Hopefully, the new FCC Chairman puts on a fast track to approval the 195 MHz option in 5 GHz and the opening up of the 10 GHz band to unlicensed use. But I’m not holding my breath. With the FCC considering the idea of reducing power output in the existing UNI-II upper band, we should all be concerned. The FCC has already handed the main course, sub 1GHz frequencies, to the big cellular companies. Next they are going to serve up the dessert, the 600 MHz bands.

There isn’t a WISP in the country or even all the WISPs together, that could get into a bidding war with the cell phone companies and the FCC knows it. So, it will be fascinating to see who objects to the 195 MHz expansion of the 5 GHz band and Hinman’s request to open up the 10 GHz band. I expect some objections from the ham radio operators up at 10 GHz band simply because that’s a legacy band for them. Brian Hinman outlined all of that in the filing and how to work with them. I’m more interested in seeing if the cellular companies, AT&T and Verizon specifically, step in and file. If that happens, and there is either a delay or outright denial of either of these bands, then we will know who runs the government (hint: it’s not the tax-paying citizens who would benefit most from these bands being opened up).

If the 10.0-10.5 GHz band is allowed to fall under the UNI-II upper rules, possibly with DFS and contention, then WISPS will have a new competitive tool. If this happens, I’m expecting big things from the manufacturing industry like Ubiquiti dropping their prices, 802.11ac equipment coming out, licensed radio manufacturers expanding their product lines, and HoHos being in stock at my local supermarket (I didn’t say it had to be radio manufacturers). Either way, I hope all of you follow the processes happening at the FCC and submit your opinions to them. The more voices that come forth, the better chance these things happen.

*Wendy’s Triple is a product of Wendy’s, the famous American hamburger chain. It consists of three burger patties stacked on top of one another and separated by thin square slices of what resembles cheese. It is considered by American healthy-eating advocates to be extremely dangerous to one’s health. Fans of Wendy’s Triple refuse to heed these warnings and ignore its calorie counts, which are clearly posted on Wendy’s website.

* * * * * * * * * * *

Previous article: Tales from the Towers, Chapter 44 – Why the Connect America Fund Should Be Abolished