Tales from the Towers Chapter 49: Ubiquiti versus Cambium, Round 2

One thing I’ve noticed is that the article comparing Ubiquiti and Cambium seems to be the most read of everything I’ve written (see Tales from the Towers Chapter 35: Ubiquiti versus Canopy/Cambium for WISPs). With the new ePMP from Cambium (and just this moment while I’m writing this, the PTP 650 announcement shows up in my email) and Ubiquiti announcing their new NanoBeam products, future 802.11ac AirPrism technologies, and the Ubiquiti World Network, I thought this would be a good time for an update. For the record though, my personal favorite was Chapter 22 (Is Law Enforcement the Red-Headed Stepchild of Broadband?) since I was able to include a ton of smarmy metaphors, Arlo Guthrie, politics, and technology. If I’d remembered to include HoHos, all the high points would have been covered.

When you stand still in the tech industry, you get run over. Unfortunately, for Motorola, they not only got run over by Ubiquiti over a period of about 3 years, they were being buried right next to Jimmy Hoffa. Right when they demonstrated the Motorola way by becoming complacent and losing an industry they created for the umpteenth time, Cambium was spun off with new ideas and motivation. Their mantra was, “this ain’t your old-school Motorola engineering team”. Even though they were managed by someone who started outside the company, all the soldiers were trained under the old regime. That meant old methodologies and ideas were going to be difficult to change. Once they had time to try and figure out how not to be Motorola, they decided the best way would be to swing the biggest stick they had, solid engineering with years of experience. They named their stick, the Avenger, or as it’s now more popularly known, the ePMP 1000.

In the meantime, Ubiquiti kept sailing the seas they launched into four years ago with the 802.11n M series radios. Over those 4 years, Ubiquiti made very few changes to the radio product line and started filling in the holes of the areas in which they had no products. The radio product line expanded a little with NanoBridges and Rockets along with new bands supplemented by a new antenna product line. The NanoBridges were a hit simply because they proved that reflectors were old school and unnecessary. The Rockets with new antennas were solid and reliable. However, the distribution channel was about as stable as the Obamacare website. Part of the problem were distributors trying to balance supply and demand. The other problems were caused by the fact that China thinks about copyright protection, patents, and ethics like bullies think about your lunch money (I could throw in human and property rights here, too, but I’m trying not to make this political. Oops, too late). If you have it and they want it, it’s now theirs.

It wasn’t all smooth sailing though. The Rocket GPS and Titaniums didn’t exactly sail, they sort of treaded water and then sank to the bottom of the ocean. There is something to be said for thoroughly testing products in the lab with the firmware before manufacturing and selling them. The old days of “ship it and let the dealer fix it” mentality that gave the US car industry such a stellar reputation in the 70s and 80s didn’t work out all that well and resulted in the Japanese taking a huge percentage of the US and world markets. The Ubiquiti variation of ship-it-and-the-firmware-will-actually-work-by-the- time-it-reaches-the-customers worked just as well. In fact, it even came with its own self-destruct AirView feature which was really handy if climbing towers multiple times to replace radios is your idea of fun. Misery loves company though, and since Titaniums still aren’t running longer than Fat Albert on a treadmill, it looks like the engineering team at Ubiquiti ran out of steam on the 802.11n product line at the Rockets without GPS. At least until 802.11ac comes out.

But meanwhile, back at the Hydro-Base . . . Remember the Avenger? This is where Cambium showed that good old American engineering talent is second to none. They not only got GPS working on an 802.11n chipset with the ePMP, but they also got more throughput out of the platform than Ubiquiti with the latest Atheros 802.11N chipset. Then they packaged it in a system that a Canopy tower-based operator would love, at least for now. Hopefully they get a little more imaginative than “if you need more range, slap a reflector on it”. Why you ask? Good question and glad you brought it up.

The Avenger, aka the ePMP, was supposed to be the Rocket and M series that Ubiquiti couldn’t build. The Avenger had to be low-cost which meant an 802.11 chipset, stable, faster, less expensive, better engineered, and a possible migratory incentive to come back for all those who strayed from the Canopy path. This was a possible resurgence to believe again in the Canopy way. No more did one have to follow the Ubiquiti disruption protocol to be able to afford starting a WISP. And no more would Cambium followers have to skip lunches because their coffers were bare when trying to start up or expand their businesses. Unfortunately, the migratory incentive got dropped in place of functionality.

While Cambium was engineering the Avenger, Ubiquiti was releasing switches, routers, AirFibers, AirGateways (which my wife thinks is the cutest little thing), potpourri, and umpteen other accessories that WISP operators would need. Unfortunately, they took their eyes off the Rocket product line which desperately needed Mariano Rivera (arguably the greatest relief pitcher ever in baseball) to take over for the original Rocket. Instead Ubiquiti tried to use the equivalent of Byung-Hyun Kim (blew back to back saves in the 2001 World Series for the Diamondbacks against the Yankees) as the Rocket GPS (if you don’t use the GPS, it works great and costs more than twice as much), and the Rocket Titanium (this one is closer to Ryan Leaf) which seems to have a shorter operational life than a fruit fly.

Back at the base, the Avenger, the 802.11 Atheros based superhero that Cambium built in the basement of Tony Stark (if Tony Stark lived in the Chicago suburbs) was released to the world. We all waited to see if it could fly in and save those of us desperate for a tower-based superhero that could fulfill the need of a more powerful 802.11n AP and toss puny Rockets off towers with real GPS functionality and throughput that would make Atheros proud. The Avenger seemed to do that and more. It’s a great product if you are starting from scratch in an area or if you want to upgrade your 5GHz Canopy products. They even added the ability to use a CPE as a base station, a la Ubquiti Nanostation, and priced the CPE under $100. The base radio and the CPE are still slightly higher than Ubiquiti but who isn’t willing to pay 10-30 percent more for a product with up to twice the throughput and solid-engineering? The ePMP is a well-designed product built by a solid engineering team and priced to make the startup WISPs think that they can finally afford Canopy. But . . and this is a heavy but . . .

Come on, you had to know after 48 articles I wasn’t going leave it at that. Cambium built a great 802.11n product that has two teensy-weensy problems. They are so minuscule that I hesitated to even mention it but I think the secret is out. Even though it’s based on an 802.11 Atheros chipset, it only speaks Canopy. In true Motorola fashion, it’s a proprietary 802.11 physical layer. It communicates with the rest of the 802.11 world the way Democrats communicate with Republicans. I don’t know if that’s going to change in the future, but I’m assuming that when they realized that they also couldn’t make GPS sync work and use an 802.11n physical layer, they decided that GPS was more important than compatibility. Man, would I have liked to have been a fly on the wall when engineering had to tell management that little tidbit during the development process. Can you imagine walking into Phil Bolt’s office to tell him that the plan to compete with Ubiquiti right in their backyard means that you only get to watch from over the fence but you still can’t play together? And your engineering team couldn’t make GPS work in 802.11 any better than Ubiquiti did?

But is that a bad thing? If your already a fully established WISP with a huge amount of Cambium investment, no. In fact, quite the contrary occurs when you can reduce your costs without taking any kind of performance penalty. If you already have an established infrastructure, this just becomes one more tool that can enhance your bottom line and can be counted on to handle growth and density. For some WISP operators that had already deployed parallel Ubiquiti equipment and found that it didn’t work out in their models, this is the answer to a prayer that doesn’t require selling your blood to afford to switch back to an all Cambium solution.

If you are a new WISP operator, this product also fits almost the same budget as a Ubiquiti deployment so startups can jump into the market with a few thousand dollars. However, if you have a Ubiquiti deployment and were hoping that Cambium would release an 802.11 AP that could provide an upgrade path for a Rocket, this isn’t going to be it. If you started with Ubiquiti, or you have a Ubiquiti deployment, then you are betting off staying with Ubiquiti and waiting to see if the 802.11ac APs from Ubiquiti will have functioning GPS.

The other minor issue is that when I last checked the calendar, 2014 was right around the corner and the bulldozers never stopped. Ubiquiti is already shipping 802.11ac radios for UniFi clients, Vivint is already field testing 4×4 802.11ac, and Xirrus will have 802.11ac by the end of November. Ubiquiti also hinted at new 802.11ac products. And to make matters even worse, they are overcoming the shortcomings of adjacent channel interference by using more advanced receiver filtering. I don’t know if Ubiquiti is going to release GPS sync in their 802.11ac products but the AirPrism idea seriously reduces the need for it.

Okay, here is where the emails and blog postings start flooding in telling me that GPS is essential and that 802.11ac has little to no value WISP deployments. I get it, believe me. If I have 1 tower covering the entire state of Oklahoma and I can get 300 frequent flyer miles to do a service call on my farthest client, 802.11ac doesn’t have much value. For GPS needs, there is the situation where 18 WISP operators on 50 towers covering farm county USA get together every Saturday night around the campfire and sing, “Imagine” (John Lennon) while synchronizing every AP between each other. Or even in an area where a single WISP has multiple towers that can see each other without a telescope. Did I mention I get it already? But did I also mention that rural WISPS cover about 10% of the WISP market. WISP operators also don’t always work together in the real world and oh yeah, RURAL WISPS ARE PROBABLY LESS THAN 10% OF THE MARKET. In almost every other case and I’m sure I’m going to get hammered on this, GPS has less value. A polling protocol is much more important.

Then there is GALACTUS, DESTROYER OF WIRED WORLDS. It isn’t going to do battle with cable and DSL providers with puny 802.11n speeds. It not only needs the faster speeds that 802.11ac is capable of executing, it also has clients that can be reached with an extended length back scratcher. DSL providers are hitting 40Mbps as long as the user can spit on the fiber switch from the front door. Cable providers are hitting 100Mbps to 1Gbps and without lining the pockets of politicians who think that it’s the taxpayers’ responsibility to make sure everyone should get NetFlix Super HD over subsidized FTTH. The only wireless difference between 802.11n and 802.11ac is about 6dBi on the s/n ratio. 802.11ac has about half the range of 802.11n or ¼ of the coverage area. If you follow the GALACTUS model, this isn’t an issue and this is the 90% of the population that WISPs should be moving into, not running away from.

To clarify this further, 2×2 802.11n with a 20MHz channel should have a realistic throughput of about 110-120Mbps. The physical layer is 144Mbps. In the real world, until the Titaniums, Ubiquiti didn’t even put 1Gbps PoE ports on the Rockets because:

(1) The processor/firmware couldn’t hit 100Mbps anyway.

(2) It would have driven the cost up another 20-30% over existing Rockets (my estimate, not theirs).

Cambium did 1Gbps ports because their APs can hit over 140Mbps (sourced from another WISP) with 40MHz wide channels. That’s an impressive statistic that means higher efficiency for Canopy tower deployments. Cambium also addressed the issue with the ePMP product line on underpowered CPEs that have plagued the Canopy 450, especially with NAT. For Canopy WISPs and new operators with short to medium range clients (at least until the SETI reflectors get in), the ePMP is a great product.

One can also argue that the ePMP with GPS also fits into the GALACTUS model perfectly. Four ePMP APs with GPS could provide up 600Mbps and even more could be added for less than the cost of a single Canopy 450. I don’t have any comparisons yet between a 450 and an ePMP in a NLOS environment with the ePMP having a proprietary physical layer, but that would be great data to know. Apparently I can also use an ePMP CPE as an AP for a few customers before needing to put in an ePMP AP. All of these are great things that I’ve been pushing for a long time. However, there are a lot of times I need to use NanoBridges in place of Nanostations because of trees and I think reflector dishes are better used for sliding down snow slopes than putting on houses (ducking from Chuck McCown).

Personally though, I’m holding out for 802.11ac. I’m looking at high densities along with higher bandwidth needs. Where the ePMP is probably around 140-150Mbps with a 40MHz channel, an 802.11ac AP is going to be around 230-250Mbps. 802.11ac processor chips are also faster, thus allowing for more functionality at the AP that is currently offloaded to routers. I’m planning for a slugfest with cable and fiber in Pleasant Valley suburbia and tying my hands behind my back with 802.11n doesn’t seem like a good idea. Vivint agrees with me apparently since they are pouring tons of money into suburban 802.11ac with a 50Mbps plan already.

One of the almost current 802.11ac products, the Xirrus 802.11ac XR-1000 products which will be available by the end of November, can also do application and security management right on the AP. Imagine being able to throttle NetFlix users, torrent users, or even virus and malware problems at the AP by application. Okay, Xirrus isn’t exactly $250, more like $800, but if I don’t have to add a Barracuda and a Procera box, it might be a bargain. And technically, it has an additional high-power Cavium processor instead of relying on an Atheros chipset so it’s really not a direct comparison. The throughput however is, and the Xirrus APs that already mash with 802.11n, won’t even blink when 802.11ac is added. At the same time, it doesn’t have GPS or polling protocols.

This article is now officially my longest one. I waited a while to absorb everything I learned at WISPAPALOOZA and see how things were shaking out with the Cambium and Ubiquiti announcements. It was an enlightening show with some great things, but the whispering around the show makes me think that not all the announcements are finished. AirPrism now has my interest piqued and I’m anxious to start testing new 802.11ac radios with it. GALACTUS isn’t used to being told to wait.

* * * * *

Previous article: Tales from the Towers, Chapter 48: Planes, Trains, Automobiles and err, Cargo Vans


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.


  1. Mike Hammett says

    I think the world already knows that we differ on your key points (AC alone will provide a greater advantage than GPS and GPS is only for rural operators), but I agree with most of everything else mentioned.

  2. Mike, what is your experience with working with other operators in the same area on synchronization? And what do you do when the other operators are running 802.11 products?

  3. I am a rural WISP. Currently a UBNT/MT shop.

    The ePMP looks interesting, but, Rocket 2M with AirPrism is a direct upgrade that would increase the capacity of my existing system… as long as the AP’s are backwards compatible with the AirMax M clients.

    While I would like GPS, I can do without it if AirPrism works… GPS failed. If AirPrism works, I should be able to go back to 3 20MHz channels on a tower in 2.4GHz. Doubling capacity of my installed base would be worth quite a bit to me…

    But if it won’t work, or if I won’t see it until 2016, I may be better off with ePMP.

  4. Mike Hammett says

    AirMax AC is NOT backwards compatible with AirMAX N. Did no one read my tweets from the show? 😉

    I’m not sure how many of us that have our Canopy systems synced as I only had to talk to one to get the settings aligned, but there are probably 8 or 10 WISPs that I’m aware of in Northern Illinois that are synced. Most of us have some pretty major coverage overlap. There are some that don’t run Canopy (I run mostly UBNT) and they’ve all been great at frequency coordination.

  5. if you use the RF Armor shields, depending on how you deploy meaning you take into account signal reflection, GPS isn’t that critical and AirPrism should pretty much eliminate that need. Kumu networks is working on some new technology also that will help too but we probably won’t see it until late 2015 at best.

  6. Where, exactly, did UBNT say that AirMax AC AP’s won’t talk to AirMax M clients?

    I’ve looked all over and can’t find that statement from UBNT.

  7. Mike Hammett says

    Ben Moore told me at the show. Their plan is similar to how they handled it with pre-AirMax clients. Eventually there will be a firmware that allows that support, but not out of the box.

  8. Wonderful. What a waste.

    Time to start looking at other gear.

  9. Blair, Mike didn’t say they wouldn’t be compatible at all, just that the won’t be compatible right when they come out. Eventually they will work together. Post this in Ubiquiti’s website and see what they say.

  10. Repost to fix typos

    Agreed, he didn’t say that.

    And, I never expected my old 802.11 UBNT gear to run with AirMax. It was a nice bonus when it did.

    I DID expect to see GPS on ALL bands as promised. I DID expect to see BeamForming on at least 2.4GHz and 5GHz. I wouldn’t have bought into AirMax except for GPS.

    A lot of my network is 802.11. some G, some N. There is some AirMax. But EVERYTHING we have can fall back to an 802.11 compatible mode. Why?

    Because I have been orphaned before. (Lucent, KarlNet, YDI, Proxim, TerraBeam, who are they this week?), and treated like a red headed stepchild. (UBNT 900MHz and 3.65GHz AirMax, RFLinx 900MHz converters)

    And I have learned from it. I am real skittish about vendor lock-in… And when it looks like that is what is happening, I tend to toss a grenade over my shoulder as I run the other way!

    I am way to small for any vendor to care about me. I know this. But, I also know that if there is no backward compatibility then why should I look to stay?

    Shipping with backward compatibility with AirMax M means that, if the improvements are as good as claimed, then it is a no brainier to upgrade my AP’s. Right now! But if it is not compatible with my existing base, it might be nice for new build-out but there is no compelling reason to replace all those CPE in the field. And, if I am only gonna use it for new build out, why not look at other options?

    At this point, all I want from UBNT in the way of new equipment is AirPrism in a Rocket M2. If it works like it they say, I’ll be all over that. I’d like to see it in 900MHz and 3.65GHz as well.

    On a side note, I still have a bit of Prism 2 and Hermes B gear in the field. In noisy environments, it still just works.

  11. Hi Rory,

    thanks for taking the time to write the article,this was the exact comparison article i was looking for.

  12. All said, I believe that epmp is a perfect product for any level of wisp. Let us not keep a blind eye on deliberant which is coming up very fast in the market with cost effective and performing devices at a fraction of Ubiquiti prices.