Tales from the Towers, Chapter 43: Galactus – Destroyer of Wired Worlds

I’ve seen projections in our industry range from “the WISP market is dying, we need to move to fiber” up to “new technologies in RF are going to make fiber the horse and buggy of bandwidth industry”.  The truth is that the wireless industry is about to put the hurt on the Wired World and they aren’t even hitting their stride yet.

My newest model which I call, Galactus – Destroyer of Wired Worlds  (say it with a low grumble in your voice for more effect), is designed just for that battle.  The good part is that you can do it today and be ready for tomorrow.  And when tomorrow comes, or really 802.11ac, it will just drop right in and re-arm you for the next few years of wired and fiber battle.  However, the WISP paradigm and deployment mentality are going to have to change to take advantage of tomorrow’s technologies.

Galactus – Destroyer of Wired Worlds – Dun, Dun, Duuunnn (we really need another Fantastic Four movie, it’s been too long), is based on the concept of a short range, low-capex, very high-capacity model.  The kind of model that can laugh in the face of wired competition knowing that nothing they are working on is going to be as competitive from a cost/benefit standpoint, at least not for the real world that most of us have to live in.  Of course, you can thank the Federal Government and its brilliant taxing policies for some of that but hey, in the US, it works to your benefit if you only provide Internet and stay away from phone service.  The reality is that with Ooma, MagicJack, and everyone having a cell phone, the POTS phone industry is on the decline.  Also, to expand into a new area, I suggest that we do what we are best at, which is deliver bandwidth.

Galactus, Galactus, Galactus – takes the idea that we now have very high-capacity, low-cost backhaul options today along with some new frequencies to work with.  What’s great is that we have decent, low-cost 802.11n equipment that can build the basic Galactus model today. Ubiquiti adding DFS frequencies to their core radios, the Rocket M5, the Nanostation M5, and the Nanobridge M5, was a huge stepping stone towards the 802.11ac market and the beginning of the Galactus model (I’m out of text-based sound effects).  It’s kind of a cross between mesh and PTMP.  Mesh needed APs every few hundred feet.  PTMP systems are designed for ranges up to several miles.  Galactus assumes low-height (sub-15 metre) APs every square kilometer or so.  Towers can be used for backhaul to the APs or some other method depending on the deployment.

With a 5.8GHz AP and 36dBm output, then you can get some serious bandwidth to a CPE to at least 1Km or more, even with vegetation. What’s amazing is that with about 50m of tree at a 1Km distance, 5.8GHz has more throughput than 2.4GHz or 3.65GHz.  Pshaw you say?  The fact is that the higher EIRP of 5GHz along with wider channels, more than compensates for the higher frequency differential and the attenuation.

With Ubiquiti 802.11n equipment, you can get somewhere around 50-100Mbps at the AP depending on what 2×2 device you are using and 50-60Mbps at the client. However, when 802.11ac comes out and you upgrade the AP, you could be talking about an AP capable of 1Gbps or more. With that kind of speed, you can also get more users per AP, even without polling protocols. DSL becomes a fading memory and your lower Capex means that Cable and Fiber companies start getting marginalized to clients that think they really need 100Mbps and are willing to pay for it.

Clients want to think they are buying the fastest service for the least amount of money even though they don’t understand it.  The consumer ideal means more is better and too much is just right, especially if it’s the cheapest.  If their video streaming doesn’t buffer and their Internet connection works when some undereducated, over-plasticized celebrity tweets some brilliant words of intellectual wisdom like “I just bought a gorgeous Gucci purse that was so infabuliciuos, I had my poodles fur dyed to match it”, they are thrilled if they can look up the purse in 1.7 seconds.  Do that, save them a few bucks, and the customers will be tripping over themselves to bail on their current provider.  802.11ac lets you advertise numbers that will blow them away, especially in markets that only have DSL providers.

Now add in the execution of 802.11ac with its ability to use wider channels than 802.11n with faster processors.  Mix it with the increased spectral efficiency and 802.11ac can theoretically handle up to 1.7Gbps at the AP with channel widths up to 160MHz.  Yeah, I know some of you are already using your fingers to make the sign of the cross yelling, NO 40MHz CHANNELS, let alone 160MHz channels.  Ha, 40MHz is for little children and weaklings and to maintain 802.11n compatibility.  I’m saying put on your big boy pants.   Anything named after a planet eater isn’t going to use puny 40MHz channels, Galactus means big.  We are talking about 80Mhz channels at least.  80MHz channels means that with 4×4 MIMO, an 802.11ac AP can kick out over 1Gbps.  And an 80MHz channel is really less than 15% of the available bandwidth up in 5GHz.  I say “80MHZ CHANNELS FOR EVERYONE, ITS ON THE HOUSE” (c’mon, you knew antenna double entendre had to be coming J).

Ahh, but what good is it if we can’t get backhaul to the AP.  Seriously, that’s the argument you are going to use?  We are wireless guys.  If there is a vertical asset, we can exploit it.  However, since Galactus was designed to be dropped right behind the enemy lines of wired infrastructure, we also need to look at everything, including the Underground Railroad, such as metro Ethernet or local fiber.  With AirFiber, Radwin, Canopy 450 (the trick here is to use the AP on both sides with one in SM mode since the SM’s aren’t fast enough)  and some really cool equipment coming out from SAF, Trango, and Exalt, there are many ways to get backhaul to local APs for very little Capex.  Even if you start out using 5.8GHz 802.11n right now (which I’m doing), you know that 802.11ac is coming out to supplement it.  For short distances from the tower, 5.4GHz can be used, for longer distance, 5.8GHz.  Both can be set up with PTMP.  Add that to the FCC promising to open up another 195MHz of bandwidth with some of it possibly using UNI-II Upper rules that means even more long range backhaul capacity.  I’m hoping that whoever takes over for Julian continues this process in the FCC.  Personally, I think the first question from the Senate during the nomination hearings should be what phone carrier the nominee uses and are they happy with it so we know who they are steering the next cellular USF subsidy to.

Finding a vertical asset every square Km or so, depending on vegetation, lets you deploy a high-capacity system that puts a WISP head-to-head with wired infrastructure.  Even though you can do it, the next question is whether not it’s financially feasible.  Well, assuming you can keep the monthly cost at about 50 percent of the revenue per AP average, it’s a smart play.  As a WISP, you can always buy your way into a market, meaning being the lowest priced provider, so there is literally no excuse for not generating $1000 per month or more off any AP.  There are a large percentage of people that are very tired of cable and DSL bills going up constantly, plus they probably hate the taxes but they don’t want to go backwards in capacity.  With 802.11ac, you can advertise DOCSIS 3.0 cable speeds and fiber bandwidth numbers at DSL prices.

So what about the 5.8GHz deployments you already have in place with clients further out than 1Km?  802.11ac is going to be better but not that much better.  The fundamental problem is that 256QAM, the highest modulation rate of 802.11ac, needs a better s/n ratio than 64QAM 802.11n. The number jumps from about 22dBm to about 28dBm to get there.  Basically to get the higher speeds of 802.11ac, you are going to get half the distance.  Canopy 450 WISPs found that out quickly when trying to move over Canopy 100 series clients since they went from FSK to OFDM.  Physics are physics.  If you are using DFS frequencies, then your distance is even shorter because of EIRP limitations.  That is both bad and good.  The good part is that it means interference is down and the s/n ratio will be very high, a perfect environment for 802.11ac.

Who knows what will happen in the future?  If the FCC gets an additional 195MHz approved and at least 100MHz of it follows UNI-II Upper rules which gives us 53dBm to play with, then a 40MHz channel still yields over 300Mbps per AP and the distances go up.  You will just be upsizing the client antennas about 6dBi over what you used to use for 802.11N or Canopy 450 deployments.  The worst case that happens is that homeowners associations get a little bit upset when you are sticking 30dBi+ antennas on roofs.

Galactus is all about getting back into the suburban trenches with wireline providers.  It’s also about future proofing to match any future speed increases from wireline.  Although DSL is toast, Intel is pushing its new 1Gb cable upgrade over DOCSIS 3.0.  However, the old adage, “Speed, Distance”, Quality” which definitely applies to 802.11ac over 802.11n, also applies to cable.  Older cable and longer distances are going to mean that a lot of infrastructure has to be replaced to get 1Gbps on cable.  Since most cable companies haven’t even amortized the cost of upgrading to DOCSIS 3.0 yet, I suspect it will be a while.  Fiber proponents are still pushing for applications that have no value for at least 10 years and no money to pay for it.  If anyone can name a single application today that can’t run over 50Mbps from a house that will advance the digitial divide or productivity in this country today, I’ll publicly apologize for telling FTTH supporters that they need to find a better cause that doesn’t involve taxpayers money.  I argue that cheaper, faster internet at 20Mbps is far more important than a transport medium that can download a Peterbuilt when the only apps we have are Smart Cars.

In the meantime, 802.11ac puts the WISP operator in the game with cable providers and GALACTUS – DESTROYER OF WIRED WORLDS (TA DA) demonstrates one model of that implementation.  But don’t wait for 802.11ac, get started today with 802.11n and be ready for it when it gets here.  Your accountant will thank you.

* * * * *

Previous article: Tales from the Towers, Chapter 42: Frustration, Thy Name is Backhaul Support

About Rory Conaway

Rory Conaway has been in the IT and Wireless Industries for the past 25 years as an author and consultant. He currently operates a growing WISP operation in Southern Arizona. He consults with investors, manufacturers, and WISPs, and develops financial business models for startups. In addition to writing articles in industry publications such as Mission Critical Magazine, Mr. Conaway also writes the series “Tales from the Towers” that can be found on various such as www.triadwireless.net and www.muniwireless.com. He has also engineered several wireless designs such as S.P.I.R.I.T. and Guerilla Wireless as well as building integrated wireless and video surveillance for airport security, municipal and critical infrastructure, SCADA systems, and hotel/MDU deployments.

Comments

  1. Mr. Conaway, very interesting read. I’m very curious about starting up a WISP in central NJ. Although there is saturation from cable/DSL/FiOS in the area, most people I’ve talked to are not very high on existing offerings, except for the triple play packages. It seems that breaking into the area while offering wireless Internet and perhaps VoIP without TV would be a hard sell. Do you see that as being the case in the areas you serve?

  2. No, not at all. You are never going to get 100% of the area. You can always buy your way into a market by being the lowest provide provider and in this economy, that’s not a hard thing to do. Considering the cost of bandwidth and overhead, that’s still a big profit. I used to think that 15% was a good starting point for market penetration but now I’m thinking more like 30% or more.

    And what possible reason do people need to get a triple play package? Pretty much everything that is on the cable TV is on Internet or over the air with few exceptions. Voice has multiple options along with all sorts of taxes and liabilities which you don’t really need to get into so point them to Ooma, MagicJack, Google, or Vonage. Bottom line, few people walk away from saving $50 a month or more unless money isn’t of concern to them . In our area, people are seeing savings of up to $130 per month when they add a Roku Box and MagicJack and that’s hard to avoid. I’ve even got customers paying 2 years in advance now.

    Do one thing and do it well, provide bandwidth.

Speak Your Mind