Mimosa MU-MIMO ?

Kick back and hang out in the lounge and talk about almost anything.
User avatar
ste
Member
 
Posts: 84
Joined: Fri May 22, 2015 5:33 am
Location: Regensburg Germany
Has thanked: 2 times
Been thanked: 13 times

Re: Mimosa MU-MIMO ?

Tue Apr 12, 2016 2:11 pm

rebelwireless wrote:I doubt most have much pre-N gear in the field these days. We have a very small number out in remote, tiny repeater sites where the equipment is still working and the throughput to the couple clients meets their plan.

One issue with the AC gear is that it's about 70% more throughput in good conditions, but anything -65 or worse and the M gear is very close, -70 and the M gear actually does better. This is fine when you are doing micro-pops (most of our new deployments) but it's focusing on one aspect that maybe wasn't that big of an issue. Chris, as you point out, you can sell high speed plans on M gear if you plan well. I would have really loved to see some efforts to improve modulation in lower SNR like Cambium (which has a 5dBm+ advantage in SNR easily) and I suspect....(NDA warning) will be much better in the short term.

That's why I'm interested in Mimosa. Yes, more throughput is great in good signal areas, but I want to be able to offer high speed service at -75 (rather, at what would be -75 on a PowerBeam to a Rocket Sector). I've already seen that 1-3dB less on Cambium on the AP & CPE ends up being the same reported signal level. Using some modern MU-MIMO and beamstearing, and a better AC radio to start with, makes me optimistic.

I'm going to go back to the fact that a Cambium 802.11n solution is VERY close to ubiquiti's AC gear on raw throughput, but is a more stable system with more predictable performance.

I can't say enough good things about M series ubiquiti, especially XM boards. Price to performance is excellent, it's really opened a lot of doors for wISPs. But their AC gear is really underwhelming and if I'm honest, should just be coming on the market now as the firmware is just getting stable enough for confident deployment. I've had AC gear hanging since the first shipment to streakwave. I've had a pretty significant script cycling through and rebooting radios for most of that time to keep everything working during normal hours.


We have some ePMP running and aren't happy at all. May be that's due to only available DFS bands here. We see dfs events, we see gps failures and we see cpe dropouts with interference. And we don't see performance comparable to UBNT AC. The epmp sm is 13 db while the nanobeam 19 is 19. There alone you get the difference to make AC work better than 11n. We had a lot of sw problems while epmp evolves and some faulty SMs.

Remember back when UBNT M was new. It was crappy like hell. So don't expect UBNT gear to work (exclude the airfibers) the first 6 months from release. Take it as showcase/beta phase. The newer beta's work quite well and perform better than UBNT M and ePMP for sure. We use the Rocket PTMP with filtering. Replacing a ePMP 1 by 1 at a 13 db Omni we see much better results. Lower latency and higher bandwidth.

User avatar
jjonsson
Associate
Associate
 
Posts: 337
Joined: Wed Nov 05, 2014 12:30 pm
Location: Denmark
Has thanked: 37 times
Been thanked: 65 times

Re: Mimosa MU-MIMO ?

Tue Apr 12, 2016 4:21 pm

rebelwireless wrote: I've had AC gear hanging since the first shipment to streakwave. I've had a pretty significant script cycling through and rebooting radios for most of that time to keep everything working during normal hours.


Actually since v6.0Beta7 / v8.0Beta7 AC equipment has been running very stable in my network. Have had no need to reboot ( > 2 months uptime).

User avatar
rebelwireless
Experienced Member
 
Posts: 607
Joined: Mon Sep 01, 2014 1:46 pm
Has thanked: 31 times
Been thanked: 136 times

Re: Mimosa MU-MIMO ?

Tue Apr 12, 2016 4:49 pm

I've stopped rebooting them in recent beta's also, but for a long time it was sketchy.

I was on early release M gear too. That's what taught me not to trust new firmware as final just because it says it's final. I was upgrading every 4-5 release after others had vetted it. I've had VERY little problems with XM because of careful firmware planning. Did have some BS with the XW hardware, but I basically only bought one shipment (and bitched thoroughly to ubnt) and kept deploying XM gear until they fixed it.

epmp vs ubiquiti is a no brainer, epmp is too slow, too basic, too low gain. I can't say that epmp even competes well with M5 hardware. performance wise, pmp450 beats the hell out of M5 gear and has a bit of an edge to AC gear in sane channel widths. Price is 2x/3x as high, but the platform is very stable today which *might* be true or almost true of ubnt AC after well over a year. When their new product comes out with beamstearing and 256QAM then ubnt AC is definitely going to be the slowest next-gen product. Cambium likes to price themselves on the very high end though so your talking $600 per SM.

Back to the topic.... Is Mimosa going to pull off MU-MIMO in a package that beats ubnt on performance and be priced similarly enough? If so, both Cambium and ubiquiti will need to react quickly.

User avatar
ste
Member
 
Posts: 84
Joined: Fri May 22, 2015 5:33 am
Location: Regensburg Germany
Has thanked: 2 times
Been thanked: 13 times

Re: Mimosa MU-MIMO ?

Tue Apr 12, 2016 5:34 pm

rebelwireless wrote:I've stopped rebooting them in recent beta's also, but for a long time it was sketchy.

I was on early release M gear too. That's what taught me not to trust new firmware as final just because it says it's final. I was upgrading every 4-5 release after others had vetted it. I've had VERY little problems with XM because of careful firmware planning. Did have some BS with the XW hardware, but I basically only bought one shipment (and bitched thoroughly to ubnt) and kept deploying XM gear until they fixed it.

epmp vs ubiquiti is a no brainer, epmp is too slow, too basic, too low gain. I can't say that epmp even competes well with M5 hardware. performance wise, pmp450 beats the hell out of M5 gear and has a bit of an edge to AC gear in sane channel widths. Price is 2x/3x as high, but the platform is very stable today which *might* be true or almost true of ubnt AC after well over a year. When their new product comes out with beamstearing and 256QAM then ubnt AC is definitely going to be the slowest next-gen product. Cambium likes to price themselves on the very high end though so your talking $600 per SM.

Back to the topic.... Is Mimosa going to pull off MU-MIMO in a package that beats ubnt on performance and be priced similarly enough? If so, both Cambium and ubiquiti will need to react quickly.


Mimosa releases A5 without TDMA without MU-MIMO and without sync. So there is time. Cambium reacted already and announced 450m which will get higher order mimo. Lets see what UBNT does.

User avatar
rebelwireless
Experienced Member
 
Posts: 607
Joined: Mon Sep 01, 2014 1:46 pm
Has thanked: 31 times
Been thanked: 136 times

Re: Mimosa MU-MIMO ?

Tue Apr 12, 2016 5:40 pm

silence from ubnt so far.

cambium's 450m is 7x MU-MIMO... (1 for control signaling in an 8x8 setup) and Mimosa's is apparently full 8x8..

seems unusual for ubnt not to pre-announce something months before hand...

User avatar
ste
Member
 
Posts: 84
Joined: Fri May 22, 2015 5:33 am
Location: Regensburg Germany
Has thanked: 2 times
Been thanked: 13 times

Re: Mimosa MU-MIMO ?

Wed Apr 13, 2016 10:03 am

rebelwireless wrote:silence from ubnt so far.

cambium's 450m is 7x MU-MIMO... (1 for control signaling in an 8x8 setup) and Mimosa's is apparently full 8x8..

seems unusual for ubnt not to pre-announce something months before hand...


Mimosa is 4x4. They show 2 A5c connected to one KP antenna but there is no possibility to make 8x8 mimo out of 2 4x4 systems. I guess they will gps-sync these 2 sectors so they could work close together and double capacity and spectrum usage. Quantenna announced a 8x8 chipset.

UBNT still needs to bring their rocket prism out of the door and implement gps sync, and improve U-CRM, and do their 2.4GHZ AC System, ...

User avatar
mike99
Associate
Associate
 
Posts: 837
Joined: Tue Nov 25, 2014 10:53 am
Location: Quebec, Canada
Has thanked: 95 times
Been thanked: 245 times

Re: Mimosa MU-MIMO ?

Fri Apr 15, 2016 9:50 am

rebelwireless wrote:cambium's 450m is 7x MU-MIMO... (1 for control signaling in an 8x8 setup) and Mimosa's is apparently full 8x8..

Do you have the source of this information ? I can't find information about 450m except that it will support beam forming + beam steering and support MU-MIMO.
rebelwireless wrote:epmp vs ubiquiti is a no brainer, epmp is too slow, too basic, too low gain.

GPS and true TDMA, not a software one, is basic for you ? The atheros chip is only for PHY and the're a second one for TDMA and GPS. Too low gain ??? Basic SM is now 16 db (eforce 180) and the're a new 25 db version (eforce 200). Soon, ePMP while also have beam forming, seem more like beam sterring to me, with ePMP 2000.

rebelwireless wrote:Mimosa is 4x4

I think it's 2 radio in 2x2, those 2 radios streams added at each other to make 4x4, so just like airfiber5 but with a 802.11 chipset. Marketing team of mimoso is as worst as Ubnt one. Exemple, picture of the A5c with fiber wire only. It must be powered by air. :roll:

User avatar
rebelwireless
Experienced Member
 
Posts: 607
Joined: Mon Sep 01, 2014 1:46 pm
Has thanked: 31 times
Been thanked: 136 times

Re: Mimosa MU-MIMO ?

Sun Apr 17, 2016 11:52 am

16dB? with a new 25dB? We've been deploying 25dBm ubiquiti radios for half a decade, with options up to 34(or ~31 typically...)
yeah, beamstearing sounds great, but higher CPE gain works too. We need high CPE gain *and* beamforming and cambium doesn't seem interested in that. For that matter, mimosa seems to be going for lower gain too. Like Cambium and Mimosa were only going for smaller coverage areas.

User avatar
jaimefink
Member
 
Posts: 5
Joined: Thu Mar 05, 2015 11:46 am
Location: San Francisco
Has thanked: 2 times
Been thanked: 2 times

Re: Mimosa MU-MIMO ?

Wed Apr 20, 2016 4:18 pm

mike99 wrote:
rebelwireless wrote:cambium's 450m is 7x MU-MIMO... (1 for control signaling in an 8x8 setup) and Mimosa's is apparently full 8x8..

Do you have the source of this information ? I can't find information about 450m except that it will support beam forming + beam steering and support MU-MIMO.
rebelwireless wrote:epmp vs ubiquiti is a no brainer, epmp is too slow, too basic, too low gain.

GPS and true TDMA, not a software one, is basic for you ? The atheros chip is only for PHY and the're a second one for TDMA and GPS. Too low gain ??? Basic SM is now 16 db (eforce 180) and the're a new 25 db version (eforce 200). Soon, ePMP while also have beam forming, seem more like beam sterring to me, with ePMP 2000.

rebelwireless wrote:Mimosa is 4x4

I think it's 2 radio in 2x2, those 2 radios streams added at each other to make 4x4, so just like airfiber5 but with a 802.11 chipset. Marketing team of mimoso is as worst as Ubnt one. Exemple, picture of the A5c with fiber wire only. It must be powered by air. :roll:


The A5c is true 4x4:4 with MU-MIMO on a single channel, with hardware GPS sync, and is fed by a single GigE PoE cable, it is not two 2x2. Read the specs first before putting bad info out there please. We've always been up front about our Backhaul radios being 2x80 MHz, with 2 RF sections fed from a 4x4 chip on 2 separate channels.

User avatar
rebelwireless
Experienced Member
 
Posts: 607
Joined: Mon Sep 01, 2014 1:46 pm
Has thanked: 31 times
Been thanked: 136 times

Re: Mimosa MU-MIMO ?

Wed Apr 20, 2016 5:39 pm

Thanks Jaime. Any chance you have some real-world examples for us? I'm eagerly waiting for something to read.... I'm testing out Cambium, but I'm coming up with it not being better enough than ubnt to justify the $$$ (for 450), or it's the epmp which is a bit back and forth and certainly not worth migrating....

I have a LOT of ubiquiti M5 out there that does 802.11n, would be VERY excited to do upgrades to A5 APs....

PreviousNext
Return to The Lounge

Who is online

Users browsing this forum: No registered users and 2 guests