Reply
Established Member
Posts: 1,672
Registered: ‎05-20-2008
Kudos: 434
Solutions: 6

Re: Gen2 vs ePMP

Yes we are missing 802.11 mode for the AC series. It has to be implemented immediately just like all WISP product vendors did.
Vendor lock-in for a 802.11 chipset doesnt make sense and is purely a business decision made by UBNT and forced upon us.
Cambium epmp1000 is expensive and outdated now. Until epmp3000 hits the shelves, there is no need to talk about Cambium.
Established Member
Posts: 1,010
Registered: ‎09-27-2015
Kudos: 494
Solutions: 17

Re: Gen2 vs ePMP


scracha wrote:

The UBNT AC stability issues have been nailed for the last year and even mixed mode has been solid for last 6 months. 


Have to disagree with this. It's getting better, but there are still customer-noticeable annoyances that haven't been fixed yet. I saw a post a couple months ago that mixed mode is scheduled for somewhat of an overhaul this year. If that's true it will probably be a lot better, but it will also probably introduce new problems.

 

Your other points are pretty spot-on. As long as UBNT doesn't turn into the Apple of wireless there isn't a compelling reason to look elsewhere.

Senior Member
Posts: 3,773
Registered: ‎04-17-2008
Kudos: 443
Solutions: 3

Re: Gen2 vs ePMP

1) Running 6x on an AC radio is just plain useless information.  I don't put up AC radios to run 64 QAM.

2) Nothing in this discussion covers the antennas.  We use Prisms with horns and because of that, we can run 40MHz channels instead of 20MHz channels.

3) Here is the part I like and I have never got an answer to.  You could put up 12 Prisms with 30 degree horns, either Ubiquiti or RF Element antennas for under $5K.  That's 12 radios that can run 40MHz channels simultaneously.  If all the clients are AC, that somewhere in the range of 600+ clients running at least 2Gbps if you spread out 50 clients exactly on each AP.  If you run half of that, imagine what kind of packages you can sell.

 

SuperUser
Posts: 5,797
Registered: ‎08-26-2009
Kudos: 1771
Solutions: 55

Re: Gen2 vs ePMP


doush wrote:

twinkletoes wrote: 

Sure, it's skewed in some ways... Just shows what the bunch of losers at Cambium are capable of after YEARS of development!


Losers ? 

 

Their 8 years old 802.11n chipset products are beating or nearly identical in performance compared to our new 802.11ac chipset products with dedicated Airmax processors. Am I missing something here ? 

 

We are an all UBNT shop here but facts are facts.


Yeah, the performance is far from identical. In several tests, limited to 6X for apples-to-apples, you can see significantly higher performance, especially for single-stream TCP. And after the next beta, you will see better still, because it can be done, it should....

SuperUser
Posts: 11,169
Registered: ‎10-28-2009
Kudos: 3760
Solutions: 320

Re: Gen2 vs ePMP


rconaway wrote:

1) Running 6x on an AC radio is just plain useless information.  I don't put up AC radios to run 64 QAM.

2) Nothing in this discussion covers the antennas.  We use Prisms with horns and because of that, we can run 40MHz channels instead of 20MHz channels.

3) Here is the part I like and I have never got an answer to.  You could put up 12 Prisms with 30 degree horns, either Ubiquiti or RF Element antennas for under $5K.  That's 12 radios that can run 40MHz channels simultaneously.  If all the clients are AC, that somewhere in the range of 600+ clients running at least 2Gbps if you spread out 50 clients exactly on each AP.  If you run half of that, imagine what kind of packages you can sell.

 


1. It really is... Even if they did only run 6x, they're still far cheaper to stock than ePMP.
2. Same here, these things fly at 40Mhz especially because of 256QAM 5/6 if you can keep CINR where it should be.
3. This is a solid point and why I used 6x 30* horns with Prisms to focus energy one direction, while co-locating with 9 other 5Ghz radios including two airFiber 5X links to other locations.  You've seen my posts about one of my Gen1 Prisms doing work at 40Mhz, 67/33, 5ms with the max of 60 subs associated and pushing well over 100Mbps at actual distances (3-6 miles) using legal power.  The thing is a < $200 radio attached to a freaking < $200 airMAX antenna, and making thousands a month in revenue for us.  I see zip point in jacking around with paying more for 'old tech' that may be solid, but doesn't offer anything more than what I'm getting now.

Tony Pierro
CTO - Wireless Internet Services, Inc.
SuperUser
Posts: 5,797
Registered: ‎08-26-2009
Kudos: 1771
Solutions: 55

Re: Gen2 vs ePMP


scracha wrote:

UBNT fanboys harping on about LTU.  Cambium fanboys (and their sales guys) harping on about UBNT not having WAVE2 but ignoring the fact Cambium doesn't currently do Wave 1.    

 

Let's stick to what's available at present.  Until they're released and stable, UBNT LTU and Cambium ePMP3000 are vaporware as far as I'm concrned.  

The only advantage of Wave 2 for WISPs is MU-MIMO...But, it is useless in many cases. GPS-timed TX synchronization plus cheap APs is MUCH more attractive than MU-MIMO.  MU-MIMO might be attractive for vendors who want to price everything sky-high, to offer some alternative to buying more expensive crap.

 

What would everyone do if UBNT offered a new 5AC airmax variant that had Wave2 chips and more chains? I don't see this being some huge improvement. Most of the sales would still go to the regular hardware. But who knows, maybe everything Cambium is thinking about for EPMP 3000 is already in alpha product form at Ubiquiti.

 


scracha wrote:

 

I'd like an 802.11 mode for UBNT AC stations as I hate vendor lock-in.  

This has been addressed here before, but obviously needs to be again.

 

The AC gear does not have an 802.11 mode and will not. It makes the software much more complicated, and requires more maintenance. It forces the entire Airmax core (which is crafted around the vendor 802.11 core) undertake major updates every time a new vendor 802.11 core is incorporated. The new 802.11 cores are important, even necessary for 802.11 compatibility, but they are detrimental to Airmax. This development effort is the primary reason the products do not incorporate 802.11 mode. As the product diverges from 802.11 even more, with Airview, Airprism, etc, this problem only gets worse.

 

Highlighted
New Member
Posts: 32
Registered: ‎09-05-2014
Kudos: 6

Re: Gen2 vs ePMP

[ Edited ]

twinkletoes wrote:

Just shows what the bunch of losers at Cambium are capable of after YEARS of development!


 Ya they must be losers to have a more stable and better performing 802.11n chipset than UBNT.

 

Here is a image of one of our ePMP2000's with Force 200 x 30 clients (new AP's going up soon for better contention, had a sales splurt here as we normally do not let this many clients per AP) I should also note the Antenna is a Rf Elements 30deg Horn.

 

20Mhz Channel with GPS 75/25 and Back to back Freq reuse enabled and in use.

 

More than happy with Cambium so far and looking forward to the ePMP3000 with AC wave 2.

(Force 300's now released and on there way to the warehouse, these will be able to connect into a ePMP2000 soon so we can atleast start getting clients on with the latest CPE while we await the ePMP3000 release)

 

We still have more UBNT in service with both M5 and AC Gen2 but are finding the Cambium more stable and performance is great on a 802.11n chipset. Can not wait to see what we will get out of the Cambium AC.

 

Looking forward to UBNT LTU though, should be great IF they develop solid firmwares.

ePMP2000.PNG
Established Member
Posts: 1,672
Registered: ‎05-20-2008
Kudos: 434
Solutions: 6

Re: Gen2 vs ePMP


twinkletoes wrote:

The AC gear does not have an 802.11 mode and will not. It makes the software much more complicated, and requires more maintenance. It forces the entire Airmax core (which is crafted around the vendor 802.11 core) undertake major updates every time a new vendor 802.11 core is incorporated. The new 802.11 cores are important, even necessary for 802.11 compatibility, but they are detrimental to Airmax. This development effort is the primary reason the products do not incorporate 802.11 mode. As the product diverges from 802.11 even more, with Airview, Airprism, etc, this problem only gets worse.

 


What are you talking about ? What development effort ? Whats complicated ?

The drivers are already there in the AC firmware. It is purely a business decision nothing else. I bet it doesnt even take 30mins for Sriram to enable the plain 802.11 mode with standard Atheros drivers. But instead UBNT is playing the vendor-lock-in game which is not the advantage of WISPs . 

Established Member
Posts: 1,036
Registered: ‎04-30-2009
Kudos: 173
Solutions: 4

Re: Gen2 vs ePMP


webformix wrote:

At least Cambium put some results together and tried to give an apples to apples comparison. THEY TRIED! What say ye UBNT? Will you actually respond to some friendly competition and "SHOW ME WHAT YOU GOT!" Or are we going to have another repeat of WISPAPALOOZA where you pull out of the competition at the last minute?

   


webformix makes a good point.  Rather than resort to childish name calling, why not post some real world testing and results.  

This thread has the typical fanboys with their typical biased opinions which does nothing for people coming on here trying to find real answers to questions and concerns they may have.  

As far as the image of wistechs results, its obvious that something isn't right.  Shows a capacity of 138 but a throughput of 148 and on top of that it shows 51% airtime.  I can guess how the fanboys will respond so you can save it  My response to that is do some real testing.

My experience.

Epmp has been more reliable from a hardware and software standpoint.  It pretty much does what it says.  The GPS works as advertised.  The support is much better.  Also, it doesn't come with alot of broken promises.  Their 2.4ghz line extremely outperforms ubiquitis.  Elevate(they made ubiquitis gear work better than ubiquiti could). The downsides are costs(for instance the gps ap is over $400), antenna offerings(limited style and gain), licensing module(inconvenient, costly). 

Ubiquiti has a diverse line of cpe and ap offerings.  They design very elegant reasonably priced equipment.  I love the prismstation and the litebeam has a price performance ratio that can't be beat.  Plus they pack their equipment in such a way that it saves a ton on shipping costs.  I'm impressed with how much they've been able to shrink container size.  Also their equipment snaps together quick and easy.  If condistions are right they can outperform on throughput by a substantion margin. The downsides are support.  Actual support is fairly limited and the forum is so frustrating.  Software has an endless number of issues(it's gotten better but still too many problems).  GPS works poorly. 2.4ghz line performs poorly.  Lots of broken promises, missed deadlines, etc.

 

It'll be interesting to see what the epmp3000 line is like once it's released.  Regardless of whether or not you like cambium , it's obvious they have pushed ubiquiti to deliver a better product.  The competiion has been good for us end users

SuperUser
Posts: 5,797
Registered: ‎08-26-2009
Kudos: 1771
Solutions: 55

Re: Gen2 vs ePMP


doush wrote:

What are you talking about ? What development effort ? Whats complicated ?

The drivers are already there in the AC firmware. It is purely a business decision nothing else. I bet it doesnt even take 30mins for Sriram to enable the plain 802.11 mode with standard Atheros drivers. But instead UBNT is playing the vendor-lock-in game which is not the advantage of WISPs . 


I hate to say this, but you have absolutely no idea what you are talking about. 

Ubiquiti Employee
Posts: 8,623
Registered: ‎04-14-2017
Kudos: 1629
Solutions: 248

Re: Gen2 vs ePMP

Have you tried GPS Sync on airMAX AC, or the newer 2AC devices at all?
SuperUser
Posts: 5,797
Registered: ‎08-26-2009
Kudos: 1771
Solutions: 55

Re: Gen2 vs ePMP


popcorrin wrote:

webformix makes a good point.  Rather than resort to childish name calling, why not post some real world testing and results.  

This thread has the typical fanboys with their typical biased opinions which does nothing for people coming on here trying to find real answers to questions and concerns they may have.  

As far as the image of wistechs results, its obvious that something isn't right.  Shows a capacity of 138 but a throughput of 148 and on top of that it shows 51% airtime.  I can guess how the fanboys will respond so you can save it  My response to that is do some real testing.

 


Thing is, 51% airtime on a 67% split can indicate a high amount of downstream usage. It's hard to perfectly utilize the downstream, especially in a software driven system like EPMP and 2AC/5AC. 


popcorrin wrote:

My experience.

Epmp has been more reliable from a hardware and software standpoint.  It pretty much does what it says.  The GPS works as advertised.  The support is much better.  Also, it doesn't come with alot of broken promises.  Their 2.4ghz line extremely outperforms ubiquitis.  Elevate(they made ubiquitis gear work better than ubiquiti could). The downsides are costs(for instance the gps ap is over $400), antenna offerings(limited style and gain), licensing module(inconvenient, costly). 

Ubiquiti has a diverse line of cpe and ap offerings.  They design very elegant reasonably priced equipment.  I love the prismstation and the litebeam has a price performance ratio that can't be beat.  Plus they pack their equipment in such a way that it saves a ton on shipping costs.  I'm impressed with how much they've been able to shrink container size.  Also their equipment snaps together quick and easy.  If condistions are right they can outperform on throughput by a substantion margin. The downsides are support.  Actual support is fairly limited and the forum is so frustrating.  Software has an endless number of issues(it's gotten better but still too many problems).  GPS works poorly. 2.4ghz line performs poorly.  Lots of broken promises, missed deadlines, etc.

 

It'll be interesting to see what the epmp3000 line is like once it's released.  Regardless of whether or not you like cambium , it's obvious they have pushed ubiquiti to deliver a better product.  The competiion has been good for us end users


Agreed, competition forces people to look further, work harder, do more...

 

Do you find that EPMP 2GHz outperforms 2AC with Prism APs? Or are you comparing M2 to EPMP ?

 

With the work going into 2AC/5AC, it has the potential to outperform EPMP in many scenarios.

 

EPMP 3000 is just catch-up to 2AC/5AC. The only benefit of this gear will be MU-MIMO. The thing is, it's worthless for existing UBNT customers who can simply plug in more APs at the same cost as the MU-MIMO Cambium unit with an extra antenna.

 

 

 

UBNT has good times coming with LTU in PTMP roles. UBNT also has good times coming with future airMAX lines...

SuperUser
Posts: 11,169
Registered: ‎10-28-2009
Kudos: 3760
Solutions: 320

Re: Gen2 vs ePMP

@popcorrin - No... it shows an AVERAGE capacity of 138Mbps (some clients are lower, some are higher). The heavy users are obviously much higher modulation to exceed the AVERAGE capacity split over 60 clients. It technically can do more than that as we have some 202Mbps capacity clients on that AP, but they weren't the ones using it when I took that snapshot. I essentially showed worst case scenario with 60 subs and it still putting in work.
Tony Pierro
CTO - Wireless Internet Services, Inc.
SuperUser
Posts: 5,797
Registered: ‎08-26-2009
Kudos: 1771
Solutions: 55

Re: Gen2 vs ePMP

heh, i never knew the AP side was an average...but now that i'm looking at my APs, it makes sense.

Regular Member
Posts: 465
Registered: ‎09-16-2013
Kudos: 149
Solutions: 14

Re: Gen2 vs ePMP

With regard to the airtime usage, this should be split up into a DL airtime usage and an UL airtime usage. That way we can set up alerts on a standard value of, say, 90% DL usage over 30 minutes in AC2. If the airtime changes with the chose UL/DL ratio, it is tough to set standard alerts. I do enjoy this feature on our Cambium deployment.
Regular Member
Posts: 336
Registered: ‎09-30-2013
Kudos: 93
Solutions: 24

Re: Gen2 vs ePMP

So check this out.  I had no idea my system could do so well. 

 

This is a Gen 2 prism Mixed Mode flex frame 20MHz with 36 clients connected to an AM-5AC21-60.  At the timeframe of this screenshot it was 28 AC clients and 8 M5 clients.  I have since replaced the M5-300 feedhorn with an AC-300.  This particular client is set to 50/5. 

 

This is off a tower in my backyard so I plan to hook up a client radio and do some tests during peak time tonight (Fiber comes into my residence and feeds the tower).  Should be interesting.  I should be able to get some realtime screenshots from the radio itself to back it up.12a busy.JPGNotice the bandwidth graph on the left. The first column is usage and the third one is capacity. The redline in this AirControl graph is the download throughput to the clients.Bob and Vicki.JPGNotice the bandwidth scales on the left. The blue line is download in Mbps which matches up with the first column. The second is signal strenghth which matches up with the yellow. The Red is Airmax Capacity %. The Aqua is the supposed link capacity which I'm leary of with the M5 gear.replaced.JPGNotice the redline dissappearing on Jan11. That is when i replaced the feedhorn of the m5 300 with an AC 300. The redline was the AirMax Capacity %. This is not a perfect client. There is a treeline along an old farm fence the signal is shooting through. That is why there is an AC-300 for a half mile shot. Right now there are no leaves on the tree. Come spring the signal will drop to -50 to -53 range.

Established Member
Posts: 1,219
Registered: ‎09-03-2013
Kudos: 311
Solutions: 37

Re: Gen2 vs ePMP


@FuzzyDicewrote:

scracha wrote:

The UBNT AC stability issues have been nailed for the last year and even mixed mode has been solid for last 6 months. 


Have to disagree with this. It's getting better, but there are still customer-noticeable annoyances that haven't been fixed yet. I saw a post a couple months ago that mixed mode is scheduled for somewhat of an overhaul this year. If that's true it will probably be a lot better, but it will also probably introduce new problems.

 

 

Ok, pretty solid :-).  I held off rollout out mixed mode for as long as possible but I've got some solar towers where it's just not cost effective to run the old Rocket M's in parallel with the AC's whilst we swap out CPEs.  In mixed mode we've had the odd nanobridge doing large latency spikes but on the whole i've been pleasantly surprised. 

 

Yeah, competition drives innovation.  I can't wait to try Cambium's ePMP3000 and also Ubiquiti's LTU (Late Tech Unicorn ?) 

Ubiquiti Employee
Posts: 8,623
Registered: ‎04-14-2017
Kudos: 1629
Solutions: 248

Re: Gen2 vs ePMP

@scracha How many M APs have you swapped out, and are you using FF mode?
Established Member
Posts: 1,219
Registered: ‎09-03-2013
Kudos: 311
Solutions: 37

Re: Gen2 vs ePMP


@UBNT-SNKwrote:
@scrachaHow many M APs have you swapped out, and are you using FF mode?

About 40 out of ~150 but we've only got about 15 in mixed mode.  The others we've either added more solar and run them in conjuntion with the M's or they've not had many subs so we've just swapped them out too.  That way if their are VOIP issues we can revert the client to the M AP.  On the busier AP's (> 20 subs) we've tended to swap out the busy (high paying!) clients and are only doing the others when they have a fault or we're doing other work in their geographical area.  

 

To be honest, the last 3 site upgrades have went so smoothly we've literally popped in the AC sectors, duplicated the M wireless

settings and then switched off the M AP's.  Netonix can switch them on remotely if required.    I think we've had 2 clients on nanobridges that had utterly bonkers latency that we just swapped out.  No real pattern why as other nanobridges off same sector same config and fine.

 

FF mode on the all AC sectors

I haven't used FF mode in mixed mode as haven't required it (for the above reasons).

Ubiquiti Employee
Posts: 8,623
Registered: ‎04-14-2017
Kudos: 1629
Solutions: 248

Re: Gen2 vs ePMP

Some NanoBridges we have seen encounter problems in some cases if they have a high TX power - I'd be interested to see what yours was on those units. Got it, that makes sense and I'm glad you're seeing some smooth transitions.
Reply