Community discussions

 
ste
Forum Guru
Forum Guru
Topic Author
Posts: 1811
Joined: Sun Feb 13, 2005 11:21 pm

802.11ac

Thu Jan 05, 2012 3:38 pm

Looks like things go further:

http://www.engadget.com/2012/01/05/broa ... s-chipset/

Promises 256QAM Modulation schemes, MuMIMO http://en.wikipedia.org/wiki/Multi-user_MIMO
and other nice stuff. As 256 QAM only operates with low noise I hope there are better filters
on these new chips.
 
User avatar
ne0031
Frequent Visitor
Frequent Visitor
Posts: 99
Joined: Tue Apr 21, 2009 10:29 pm

Re: 802.11ac

Thu Jan 05, 2012 4:40 pm

Given the pain we have all dealt with as we have moved into n, I doubt we'll see any perceptible traction with Mikrotik support any time soon.
Matt

ABC/WXYZ - you get the picture.....

150khz-110ghz
1/5sps to 4gbps
 
ste
Forum Guru
Forum Guru
Topic Author
Posts: 1811
Joined: Sun Feb 13, 2005 11:21 pm

Re: 802.11ac

Thu Jan 05, 2012 6:52 pm

Given the pain we have all dealt with as we have moved into n, I doubt we'll see any perceptible traction with Mikrotik support any time soon.
Sure. But as n is going to mature, it will be with 802.11ac. At least it gives some prospect to the
future. As it will be in every laptop very soon the market will give us hightech for a very low price
again.
 
0ldman
Forum Guru
Forum Guru
Posts: 1446
Joined: Thu Jul 27, 2006 5:01 am

Re: 802.11ac

Thu Jan 05, 2012 7:21 pm

Last time I read up on 802.11ac it sounded less like an outline of a plan and more like a bunch of excited kids deciding what their first car is going to be like.

"I want it to be fast"
"I want it to be better against interference"
"I want it to use larger channels"
"I want it to use rainbows and unicorns"

The local cable company changed over to 256QAM recently and it was very difficult, took a year or more before they got the cable plant where it needed to be, and they are shielded from most interference.

I think 802.11ac will cut about half of the proposed features out and blow 802.11n out of the water, but I don't think it will progress until they stop trying to do the impossible.
 
ste
Forum Guru
Forum Guru
Topic Author
Posts: 1811
Joined: Sun Feb 13, 2005 11:21 pm

Re: 802.11ac

Thu Jan 05, 2012 7:51 pm

Last time I read up on 802.11ac it sounded less like an outline of a plan and more like a bunch of excited kids deciding what their first car is going to be like.

"I want it to be fast"
"I want it to be better against interference"
"I want it to use larger channels"
"I want it to use rainbows and unicorns"

The local cable company changed over to 256QAM recently and it was very difficult, took a year or more before they got the cable plant where it needed to be, and they are shielded from most interference.

I think 802.11ac will cut about half of the proposed features out and blow 802.11n out of the water, but I don't think it will progress until they stop trying to do the impossible.
We use 256QAM for years. It's standard for licensed gear and works great. Motorola PTP600 does higher modulations
in 5GHz for years giving good results (If you've the money). Of course it needs better conditions than
64QAM but for short or clear links it gives a boost. Let's see how Broadcom/Atheros does the job.
 
0ldman
Forum Guru
Forum Guru
Posts: 1446
Joined: Thu Jul 27, 2006 5:01 am

Re: 802.11ac

Thu Jan 05, 2012 8:11 pm

I am very interested in what they can do with it.

I guess its one of those "reach for the stars and pull back as necessary" plans. Not a bad thing, but I prefer to under sell and over deliver.

I still have issues with 54mb ptmp in a fairly rural area.
 
syadnom
Member
Member
Posts: 405
Joined: Thu Jan 27, 2011 7:29 am

Re: 802.11ac

Sun Jan 22, 2012 7:57 pm

the 256QAM modulation will be great for me. I run backhaul links between .5 miles and 5 miles with VERY good SNR so I suspect that 256QAM wont be a problem for me.

I don't really expect to deliver to the average customer at this modulation considering I frequently deliver at just 16QAM modulation with excellent results. That said, I can deliver 64QAM to a huge number of subscribers today but hard setting the rate on the radios gives more predictable performance (no latency spikes when reaching the max throughput for the given link for example)
 
mjj
just joined
Posts: 17
Joined: Sat Jun 16, 2007 3:55 pm

Qualcomm Atheros is ready?

Sun Dec 02, 2012 1:11 am

Qualcomm Atheros have QCA9890 3x3 and QCA9892 2x2 802.11ac chipsets and "Qualcomm Atheros is sampling its 802.11ac chips to customers in the second quarter of 2012...".

So... when we should have info from MT team about compatibility & new cards from Mikrotik?
 
Lakis
Forum Veteran
Forum Veteran
Posts: 701
Joined: Wed Sep 23, 2009 7:52 pm

Re: 802.11ac

Sun Dec 02, 2012 3:20 pm

"I want it to be better against interference"
Commercial crap
interference is interference u cant be better against interference it is there in the air
and the question is how to avoid interference
 
0ldman
Forum Guru
Forum Guru
Posts: 1446
Joined: Thu Jul 27, 2006 5:01 am

Re: 802.11ac

Wed Dec 05, 2012 8:19 pm

"I want it to be better against interference"
Commercial crap
interference is interference u cant be better against interference it is there in the air
and the question is how to avoid interference
That was my entire point.

802.11ac isn't going to end up being the savior of WISPs. If anything it sounds like it has more liabilities.
 
syadnom
Member
Member
Posts: 405
Joined: Thu Jan 27, 2011 7:29 am

Re: 802.11ac

Wed Dec 05, 2012 10:51 pm

"I want it to be better against interference"
Commercial crap
interference is interference u cant be better against interference it is there in the air
and the question is how to avoid interference
That was my entire point.

802.11ac isn't going to end up being the savior of WISPs. If anything it sounds like it has more liabilities.
0ldman, I'm not really arguing, but what specifically are you thinking will be liabilities? Are you speaking of other device interference, or other wisps that are irresponsible with spectrum?

ac's 8X8 MIMO capabilities is like 3/4 of beamforming. It doesn't steer the beam by timing tx on multiple elements, but does tx on multiple elements in slightly different directions/polarities which can improve SNR on the client side. Also, with 8 receiving elements and algorithms to identify desired transmitions it can avoid noise and even filter it. a -3dB improvement on the client and on the AP side is a very nice gain.

Now, 80Mhz channels in ISM band is probably going to cause some grief...
 
Basiley
Member Candidate
Member Candidate
Posts: 101
Joined: Thu Dec 06, 2012 2:42 pm

Re: 802.11ac

Tue Dec 11, 2012 9:39 pm

cool. how about 802.11ad chips too in RB then ? they complimentary for ac[just 60GHz derrivative], if im read specs/standard well/correct.
its not something bogus/narrrow-demanded, like 802.22 =)
or widely used but irrelevant yet[like ZigBee]to networking -)
Last edited by Basiley on Thu Dec 13, 2012 11:03 am, edited 1 time in total.
 
syadnom
Member
Member
Posts: 405
Joined: Thu Jan 27, 2011 7:29 am

Re: 802.11ac

Tue Dec 11, 2012 10:06 pm

cool. how about 802.3ad chips too in RB then ? they complimentary for ac[just 60GHz derrivative], if im read specs/standard well/correct.
its not something bogus/narrrow-demanded, like 802.22 =)
or widely used but irrelevant yet[like ZigBee]to networking -)
802.3ad doesn't requre radio support and can be run across two wireless links now. No special 802.11ac+.3ad work is required.
 
Basiley
Member Candidate
Member Candidate
Posts: 101
Joined: Thu Dec 06, 2012 2:42 pm

Re: 802.11ac

Thu Dec 13, 2012 11:04 am

cool. how about 802.3ad chips too in RB then ? they complimentary for ac[just 60GHz derrivative], if im read specs/standard well/correct.
its not something bogus/narrrow-demanded, like 802.22 =)
or widely used but irrelevant yet[like ZigBee]to networking -)
802.3ad doesn't requre radio support and can be run across two wireless links now. No special 802.11ac+.3ad work is required.
oh, sorry, my typo. 802.11ad meant :(
http://www.ieee802.org/11/Reports/tgad_update.htm
that wouldn't be[because air absorption and nearly-zero wall-penetratio]any interests of ISP[significant portion of MikroTik consumers], but would be quite attractive for home consumers, corporations and integrators.
 
syadnom
Member
Member
Posts: 405
Joined: Thu Jan 27, 2011 7:29 am

Re: 802.11ac

Thu Dec 13, 2012 6:32 pm

Ah, that makes MUCH more sense :)

60Ghz isn't so bad for short hops. Current EIRP is 40dBm and 500mW radio. There is a proposal in the FCC to take this to 82dBm EIRP when the antennas is >=51dBi.

The curious thing to me about 'ad' is that it is simultaneous tri-band of 2.4,5,60Ghz. For WISP service, this is a short range backhaul with integrated 2.4 and 5Ghz backups. 60Ghz has a tiny fresnel which is nice to work with.

The 'catch' with 60Ghz is that it is attenuated more by O2. This isn't so bad though because O2 levels remain pretty consistent through weather and seasons. Unlike frequencies that attenuate significantly in H2O (2.4, 24), 60Ghz links should be more consistent. I recall reading an anecdote somewhere that 60-90Ghz perform better in fog or snow because the moisture displaces O2, but I can't find that on google atm.
 
Basiley
Member Candidate
Member Candidate
Posts: 101
Joined: Thu Dec 06, 2012 2:42 pm

Re: 802.11ac

Fri Jan 11, 2013 5:19 am

Ah, that makes MUCH more sense :)

60Ghz isn't so bad for short hops. Current EIRP is 40dBm and 500mW radio. There is a proposal in the FCC to take this to 82dBm EIRP when the antennas is >=51dBi.

The curious thing to me about 'ad' is that it is simultaneous tri-band of 2.4,5,60Ghz. For WISP service, this is a short range backhaul with integrated 2.4 and 5Ghz backups. 60Ghz has a tiny fresnel which is nice to work with.

The 'catch' with 60Ghz is that it is attenuated more by O2. This isn't so bad though because O2 levels remain pretty consistent through weather and seasons. Unlike frequencies that attenuate significantly in H2O (2.4, 24), 60Ghz links should be more consistent. I recall reading an anecdote somewhere that 60-90Ghz perform better in fog or snow because the moisture displaces O2, but I can't find that on google atm.
presently only SoC horsepower demands throttling 60Ghz band market penetration[ie cost of chips], just like how 10G-Base-T halt in advances last 6 years, depsite CAT6a and CAT7+ cables/connectors and chips improvements.
similar issues - prevent creating affordable holo-displays and holo-storage, too, for reference, and portable/operative DNA pocket sequencers/verifiers.
 
rado3105
Member
Member
Posts: 480
Joined: Sat Jan 12, 2008 11:45 pm

Re: 802.11ac

Fri Jan 11, 2013 7:06 pm

0ldman, I'm not really arguing, but what specifically are you thinking will be liabilities? Are you speaking of other device interference, or other wisps that are irresponsible with spectrum?

ac's 8X8 MIMO capabilities is like 3/4 of beamforming. It doesn't steer the beam by timing tx on multiple elements, but does tx on multiple elements in slightly different directions/polarities which can improve SNR on the client side. Also, with 8 receiving elements and algorithms to identify desired transmitions it can avoid noise and even filter it. a -3dB improvement on the client and on the AP side is a very nice gain.

Now, 80Mhz channels in ISM band is probably going to cause some grief...
yes this is right, the problem is that many people doesnt understand what really is 802.11ac. Beamforming(later revisions I think) and MuMIMO(later revisions) are the most expected. There more spatial streams on both sites, and because of better characteristic of this spatial streams, using computer algorithms it can be more effectively filtered noise. Because more streams on ap site, more streams or client site..and both sites can better know the enviroment - what is noise, and what is wanted traffic. On this are based other competitors like Ruckus, Cisco...that have working solutions....
Also this can lead to adaptive filtering of enviroment...Because radios are built for various enviroment, and every enviroment is different. But when some radio is in one enviroment this enviroment is pretty stable....so radio can adapt to it.....
 
syadnom
Member
Member
Posts: 405
Joined: Thu Jan 27, 2011 7:29 am

Re: 802.11ac

Fri Jan 11, 2013 7:23 pm

rado3105, this brings up a good point about 'ac. It's not beamforming, but it accomplishes some similar things.

It's more 'beamselection'. The receive elements are all individually capable of tx and rx. The radio may determine that client x is most visible (better signal, SNR, or lack of echos or noise) on elements 2, 4, 7 and can then chose to ignore or lessen the impact of signals that look like client x on the other elements. This, in a multi-user MIMO setup, would allow for client y to be communicated with on some other combination of elements, 1,2,5. If these two clients are either near each other, or in such a position that some refraction or echo makes them look like they are near each other on some elements, the radio can scrap the signal received on the less than optimal elements.

just to illustrate why 'ac isn't beamforming, 'ac will transmit on all elements it chooses to at the same time. AP > client x would tx on 2, 4, 7 (or whatever was learned to be best for the client) at the same time. beamforming would delay the tx on certain elements to cause a wavefront peak at the client like two waves in a pond colliding and aggregating their power. The delay determining the aggregation points. 'ac and beamforming will handle rx almost identically because the beamforming transmitter handled calculating the delays, the receiver should get symbols from the elements synchronously.

So, a beamforming AP doesn't really need a beamforming client for async communications, though the client would need to be aware of the AP in the sense that it should reply back with the timing difference of the beacons used for beamforming. That's really a firmware thing though. A beamforming client would be an improvement though as it would be able to tx with beamforming as well as return better data to the AP about beacon strength as well as timing. Basically faster, more accurate timing and element selection.
 
rado3105
Member
Member
Posts: 480
Joined: Sat Jan 12, 2008 11:45 pm

Re: 802.11ac

Fri Jan 11, 2013 7:56 pm

rado3105, this brings up a good point about 'ac. It's not beamforming, but it accomplishes some similar things.

It's more 'beamselection'. The receive elements are all individually capable of tx and rx. The radio may determine that client x is most visible (better signal, SNR, or lack of echos or noise) on elements 2, 4, 7 and can then chose to ignore or lessen the impact of signals that look like client x on the other elements. This, in a multi-user MIMO setup, would allow for client y to be communicated with on some other combination of elements, 1,2,5. If these two clients are either near each other, or in such a position that some refraction or echo makes them look like they are near each other on some elements, the radio can scrap the signal received on the less than optimal elements.

just to illustrate why 'ac isn't beamforming, 'ac will transmit on all elements it chooses to at the same time. AP > client x would tx on 2, 4, 7 (or whatever was learned to be best for the client) at the same time. beamforming would delay the tx on certain elements to cause a wavefront peak at the client like two waves in a pond colliding and aggregating their power. The delay determining the aggregation points. 'ac and beamforming will handle rx almost identically because the beamforming transmitter handled calculating the delays, the receiver should get symbols from the elements synchronously.

So, a beamforming AP doesn't really need a beamforming client for async communications, though the client would need to be aware of the AP in the sense that it should reply back with the timing difference of the beacons used for beamforming. That's really a firmware thing though. A beamforming client would be an improvement though as it would be able to tx with beamforming as well as return better data to the AP about beacon strength as well as timing. Basically faster, more accurate timing and element selection.
I was primarly writing about PtP....you mainly about PtM
 
syadnom
Member
Member
Posts: 405
Joined: Thu Jan 27, 2011 7:29 am

Re: 802.11ac

Fri Jan 11, 2013 8:02 pm

I was primarly writing about PtP....you mainly about PtM
It's effectively the same as far the this conversation goes. minus MU-MIMO. It's PtMP with only 1 client. I doubt there will be many efforts to create a specific PtP type 'ac connection, it will be just like nv2/airmax/802.11 are today, a PtP has an AP and a client.
 
ste
Forum Guru
Forum Guru
Topic Author
Posts: 1811
Joined: Sun Feb 13, 2005 11:21 pm

Re: 802.11ac

Mon Dec 09, 2013 10:38 am

Hi MT,

Holiday season is showing up soon. Some notes from your wireless development lab
would lighten up years end :) .

So what's up with 802.11ac. Will we see first samples of ac early next year?

Stefan
 
0ldman
Forum Guru
Forum Guru
Posts: 1446
Joined: Thu Jul 27, 2006 5:01 am

Re: 802.11ac

Tue Dec 17, 2013 6:33 am


0ldman, I'm not really arguing, but what specifically are you thinking will be liabilities? Are you speaking of other device interference, or other wisps that are irresponsible with spectrum?

ac's 8X8 MIMO capabilities is like 3/4 of beamforming. It doesn't steer the beam by timing tx on multiple elements, but does tx on multiple elements in slightly different directions/polarities which can improve SNR on the client side. Also, with 8 receiving elements and algorithms to identify desired transmitions it can avoid noise and even filter it. a -3dB improvement on the client and on the AP side is a very nice gain.

Now, 80Mhz channels in ISM band is probably going to cause some grief...
256QAM has potential but likely to have problems with interference. I see it having to be unchecked to have a stable link in most PTMP situations. I hope I'm wrong. I could use the bandwidth.
80MHz wide channel is going to be harder to keep clean. Hopefully the FCC will give us some more bandwidth, but I'm not holding my breath. UNII 3 is what, 100MHz wide?

It isn't just other WISP, all sorts of devices are using 5GHz, even something to do with the auto braking systems on cars.

I am interested to see how the multiple antenna thing will handle interference, but so far I haven't read where someone has even hacked a consumer router for outdoor use, still haven't seen anything with 8 antennas yet either.

802.11n can handle 4 or 6 chains max? How many have you seen that actually use 802.11n to its fullest?
 
syadnom
Member
Member
Posts: 405
Joined: Thu Jan 27, 2011 7:29 am

Re: 802.11ac

Tue Dec 17, 2013 6:57 am

4 chains max.

802.11n just didn't have enough in the spec to handle that many chains in most environments. I've seen a number of 3 chain setups that worked nicely. I the 4th chain just adds too much noise for that rather weak 'n radio to handle.

I'm running unifi-ac aps with 6x6 and it is great. I've used a dlink 'ac unit that was faster but didn't have a stable of connection.

I'm sold on big fat channels and lots of them for indoor use. It works very well and I can pull in really fenominal speeds.

For outdoor, I'm still convinced that 80Mhz is a absolute waste, leaving none of the preferred CPE band available. Also, I can't imagine getting 256QAM off of consumer radios. That said, I have played with the ubiquiti outdoor unit that has the prism chip, that thing is killer! If they can put a prism in front of 'ac radios, then maybe 256QAM isn't far fetched.
 
ste
Forum Guru
Forum Guru
Topic Author
Posts: 1811
Joined: Sun Feb 13, 2005 11:21 pm

Re: 802.11ac

Tue Dec 17, 2013 11:17 am

256QAM has potential but likely to have problems with interference. I see it having to be unchecked to have a stable link in most PTMP situations. I hope I'm wrong. I could use the bandwidth.
80MHz wide channel is going to be harder to keep clean. Hopefully the FCC will give us some more bandwidth, but I'm not holding my breath. UNII 3 is what, 100MHz wide?
I guess 256QAM will only work in smaller channels or on PTP with high gain antennas.
But even then it will have it's benefits. With small size installations we see a lot of customers
on highest 64QAM modulations. These may Switch to 256QAM and therefore increase the sector capacity.

My hope is that implementing 256QAM and bigger channel forces Atheros/Broacom to build higher quality radios which have a better co-channel interference handling.
 
0ldman
Forum Guru
Forum Guru
Posts: 1446
Joined: Thu Jul 27, 2006 5:01 am

Re: 802.11ac

Thu Jan 23, 2014 1:02 am

I haven't confirmed this yet, but rumor has it that Ubiquiti has cleaned up their signal quite a bit with software only. Going from 5.3.5 to 5.5.4 made the off channel drop off steeper, co-location easier and channel overlap cleaner, less of an issue.

I haven't checked MT's stuff since updating to 5.25, but with the results I've had after upgrading follow suit with the UBNT rumor. I get cleaner signals on the same 433 board with 3 R52H running 2.4GHz than I ever had with 3.30. I think these guys can pull some pretty nice stuff off in software if they focus on it. I know a lot of that comes down to what LNA they use.

Still interested in 802.11ac, google search is what brought me back to this thread. I'm upgrading my main backhaul soon, some MT 802.11ac hardware would be awesome.
 
syadnom
Member
Member
Posts: 405
Joined: Thu Jan 27, 2011 7:29 am

Re: 802.11ac

Thu Jan 23, 2014 7:01 am

I haven't confirmed this yet, but rumor has it that Ubiquiti has cleaned up their signal quite a bit with software only. Going from 5.3.5 to 5.5.4 made the off channel drop off steeper, co-location easier and channel overlap cleaner, less of an issue.
Maybe a small improvement. Nothing miraculous.

Still interested in 802.11ac, google search is what brought me back to this thread. I'm upgrading my main backhaul soon, some MT 802.11ac hardware would be awesome.
What do you use for backhaul now? AirFiber5 and 24 are pretty awesome!
 
server8
Member
Member
Posts: 396
Joined: Fri Apr 22, 2011 1:27 pm

Re: 802.11ac

Thu Jan 23, 2014 10:34 am

I haven't confirmed this yet, but rumor has it that Ubiquiti has cleaned up their signal quite a bit with software only. Going from 5.3.5 to 5.5.4 made the off channel drop off steeper, co-location easier and channel overlap cleaner, less of an issue.
I tested in the lab AIRPRISM@2.4 Ghz I am not sure but I think it is a trik they use a little bit smaller channel width.

I also tested in the lab 802.11AC (compex HW with ATH10K) , QAM256 (2 chains) is impressive 30% faster. I have no idea how it work in an open and crowled enviroment I hope to test it soon.

Giuseppe
 
Tbird
Frequent Visitor
Frequent Visitor
Posts: 55
Joined: Tue Dec 09, 2008 8:32 pm

Re: 802.11ac

Thu Jan 23, 2014 11:39 am

We have been running a couple of 25dBi AC PtP links in Sweden for 6 months now - and it is awesome! To take full advantage of the 256QAM and 5/6 coding rate, you obviously need good signal - but you also need really good port to port isolation on the antennas.

On 80MHz channels we see close to 600 Mbps throughput on the shorter links - at 867 Mbps link rate. RFC2544 tests showed 180 Mbps full duplex.

Our experience is to avoid any 3x3 MIMO antennas, at least for PtP - simply not polarity separation enough to work properly for any distance at all.

/TB
 
mp3turbo
just joined
Posts: 23
Joined: Fri May 29, 2009 9:24 pm

Re: 802.11ac

Thu Jan 23, 2014 12:37 pm

Phenomenal results !! What devices do you use ?
 
Tbird
Frequent Visitor
Frequent Visitor
Posts: 55
Joined: Tue Dec 09, 2008 8:32 pm

Re: 802.11ac

Thu Jan 23, 2014 3:36 pm

This most of all due to good antenna, with -45 to -50 dB isolation, and integrated with aluminum enclosure for good collocation noise suppression. On one location there are more than 20 other APs "visible" - still good performance, but this is naturally hard to quantify.

Another surprise on AC was the "multi MCS" support - you may have CPEs connected at a mix different channel sizes - simultaneously! You may set AP in 20/40/80 MHz mode, and you can connect CPEs at any mix of these. Have seen other tests claim 40% better performance on N-clients when using AC as AP, but we have no experience to confirm this aspect yet.

Mikrotik forum - so cannot name brands, but hopefully Mikrotik will support AC soon also.

/TB
 
ste
Forum Guru
Forum Guru
Topic Author
Posts: 1811
Joined: Sun Feb 13, 2005 11:21 pm

Re: 802.11ac

Thu Jan 23, 2014 4:05 pm

This most of all due to good antenna, with -45 to -50 dB isolation, and integrated with aluminum enclosure for good collocation noise suppression. On one location there are more than 20 other APs "visible" - still good performance, but this is naturally hard to quantify.

Another surprise on AC was the "multi MCS" support - you may have CPEs connected at a mix different channel sizes - simultaneously! You may set AP in 20/40/80 MHz mode, and you can connect CPEs at any mix of these.
Isn't it the same as in 11n? You have a main 20MHz Channel and an Extension channel (above or below). Only CPEs with good signal use the extension channel.
 
syadnom
Member
Member
Posts: 405
Joined: Thu Jan 27, 2011 7:29 am

Re: 802.11ac

Thu Jan 23, 2014 7:17 pm

This most of all due to good antenna, with -45 to -50 dB isolation, and integrated with aluminum enclosure for good collocation noise suppression. On one location there are more than 20 other APs "visible" - still good performance, but this is naturally hard to quantify.

Another surprise on AC was the "multi MCS" support - you may have CPEs connected at a mix different channel sizes - simultaneously! You may set AP in 20/40/80 MHz mode, and you can connect CPEs at any mix of these.
Isn't it the same as in 11n? You have a main 20MHz Channel and an Extension channel (above or below). Only CPEs with good signal use the extension channel.
In 'n it's exactly as you say, regular clients use 20, 40Mhz clients can use the extension. 'n modulates the whole channel, 1 MCS12 user brings the AP to MCS12

In 'ac, it can modulate each sub-carrier differently.

In 'n and 'ac (not 'g) there are 56 sub-carriers in 20Mhz, so 'ac being able to to isolate up to 56 separate sub-carriers with different MCS values allows those low-performing APs to get modulated down on just the sub-channels that AP assigns to them and clients with good signal can get better speeds.
 
Tbird
Frequent Visitor
Frequent Visitor
Posts: 55
Joined: Tue Dec 09, 2008 8:32 pm

Re: 802.11ac

Fri Jan 24, 2014 12:08 pm

Right on! - Aruba have some nice in-depth white paper on multi-user MIMO and modulation - for those who wants' to expand their understanding.

I've also seen some simulation results concluding that 100 subscribers is no problem for AC. Hmm - we definitely need to test out AC in PtMP mode here now.

/TB
 
syadnom
Member
Member
Posts: 405
Joined: Thu Jan 27, 2011 7:29 am

Re: 802.11ac

Fri Jan 24, 2014 11:14 pm

Here is about the best 'ac write-up that I have seen.
http://www.cisco.com/en/US/prod/collate ... 13103.html

As far as 100 subs, 'ac will have the same loss of throughput as 'n where a lot time is spent just maintaining the link to the client, but the scheduling mechanism is much better, having up to 8 chains can help isolate clients and reduce noise, and 'ac chips will simply have faster cpus which will help.

I don't think that 100 subs would be realistic for wISP services and next-gen bandwidth needs. 'ac is not REALLY offering much more bandwidth than 'n, it's offering wider channels, more spacial streams, more cpu power, and better scheduling. If you are currently running 'tik or ubnt APs with nv2/airmax you can probably pull off 50 clients w/ 3-6Mb service but NOT 9-12Mb service. 'ac will likely make 9Mb x50 clients much better but without eating up ALL of your spectrum on 1 AP, you aren't going to have enough bandwidth to have more clients. Additionally, I don't think you will get enough clients on a single AP to just eat up a bunch of spectrum with a big fat channel.

It would be slick if you could just have one 80Mhz channel and get 250 subs, but that's not realistic :/
 
Tbird
Frequent Visitor
Frequent Visitor
Posts: 55
Joined: Tue Dec 09, 2008 8:32 pm

Re: 802.11ac

Sat Jan 25, 2014 12:41 am

Yeah, grate write-up by Cisco - we are definitely still learning here!

Totally agree - 100 CPEs is not realistic at all, and these simulations probably don't consider much frame-loss and retransmitting that will occur in the wild - causing scheduling issues . Still, if you could get half of that it would still be a huge improvement.

We did some comparing testing on raw data throughput for comparing N with AC in PtP (RFC2544):
@20MHz - 110Mbps vs 150 Mbps
@40MHz - 260Mbps vs 300 Mbps
@80MHz - NA vs 580 Mbps

So there are some gain from AC for sure, but no miracles. To take full advantage of the 256QAM and 5/6 coding rate, we see that good signals and good antennas with proper port to port isolation, is more important than ever.

/TB
 
Lakis
Forum Veteran
Forum Veteran
Posts: 701
Joined: Wed Sep 23, 2009 7:52 pm

Re: 802.11ac

Sat Jan 25, 2014 4:00 am

so we come to ac standard do you all understated what is ac - standard that provide up to 1.5 Gbit/s
no home desktop have cpu that can process such 1.5 Gbit per second information
no mikrotik board can provide that kind of speed

Sumsung S4 have Wi-Fi 802.11 a/b/g/n/ac - c'mon ac on phone why? There is no known internet connection that can provide 1.5gbit/s

admitted u all need ac for fast p2p links.. like airfiber :)
 
ste
Forum Guru
Forum Guru
Topic Author
Posts: 1811
Joined: Sun Feb 13, 2005 11:21 pm

Re: 802.11ac

Sat Jan 25, 2014 1:58 pm

so we come to ac standard do you all understated what is ac - standard that provide up to 1.5 Gbit/s
no home desktop have cpu that can process such 1.5 Gbit per second information
no mikrotik board can provide that kind of speed

Sumsung S4 have Wi-Fi 802.11 a/b/g/n/ac - c'mon ac on phone why? There is no known internet connection that can provide 1.5gbit/s

admitted u all need ac for fast p2p links.. like airfiber :)
Looking at 11n you never see the advertised speeds. If 11ac increase speed I am sure it is usable.
 
Lakis
Forum Veteran
Forum Veteran
Posts: 701
Joined: Wed Sep 23, 2009 7:52 pm

Re: 802.11ac

Sun Jan 26, 2014 1:38 am

ac
wider RF bandwidth (up to 160 MHz) - this is bad this will bring much much more interference in air
for home usage its ok
Support for up to eight spatial streams - this is good this is how should look future of wireless more and more steams + it can have up to 16 antennas one for rx one for tx
 
ste
Forum Guru
Forum Guru
Topic Author
Posts: 1811
Joined: Sun Feb 13, 2005 11:21 pm

Re: AW: 802.11ac

Sun Jan 26, 2014 11:38 am

ac
wider RF bandwidth (up to 160 MHz) - this is bad this will bring much much more interference in air
for home usage its ok
Support for up to eight spatial streams - this is good this is how should look future of wireless more and more steams + it can have up to 16 antennas one for rx one for tx
We use only 20 MHz with 11n in most sites but 11ac has more benefits than bigger channels.

Gesendet von meinem LG-D802 mit Tapatalk
 
vladimirslk
Member Candidate
Member Candidate
Posts: 109
Joined: Wed Feb 10, 2010 2:03 am
Location: Estonia, Tallinn
Contact:

Re: 802.11ac

Thu Feb 06, 2014 9:07 am

ubiquiti have AC devices since september I suppose, check their site.
for AP price is ca 250$
 
User avatar
normis
MikroTik Support
MikroTik Support
Posts: 24259
Joined: Fri May 28, 2004 11:04 am
Location: Riga, Latvia

Re: 802.11ac

Thu Feb 06, 2014 9:10 am

ubiquiti have AC devices since september I suppose, check their site.
for AP price is ca 250$
It's not really an AP, it's their unify system.
No answer to your question? How to write posts
 
Beccara
Long time Member
Long time Member
Posts: 606
Joined: Fri Apr 08, 2005 3:13 am

Re: 802.11ac

Thu Feb 06, 2014 11:24 pm

ubiquiti have AC devices since september I suppose, check their site.
for AP price is ca 250$
It's not really an AP, it's their unify system.
I'm connected to one with nothing installed on any servers or anything, it conf'ed up like a normal AP and It's providing me access from a point
 
wispwest
Member
Member
Posts: 477
Joined: Tue May 19, 2009 3:48 am

Re: 802.11ac

Sat Feb 08, 2014 6:15 am

people are so worked up about how much spectrum AC hogs up. what I'm excited about, is the conservative side of it. I believe 20mhz on 802.11ac will give more throughput than 20mhz on 802.11n.
 
wispwest
Member
Member
Posts: 477
Joined: Tue May 19, 2009 3:48 am

Re: 802.11ac

Sat Feb 15, 2014 8:15 am

Is there ANY update on this, or is Mikrotik going to be left in the dust??

I talked to some people that are beta-testing the new Ubiquiti Airmax-AC products. It's unbelievable! Even on smaller channelwidths they're seeing like 130mb on 20mhz (air rates 177mb) 230mb on 40mhz channel width (air rates 400mb) and up to 500mb throughput on 80mhz channel width (866mb air rates).

Not sure when they'll officially release it, but it sounds like even the beta products being tested are going quite smooth. These people signed NDA's, so not much else is known other than this stuff is really revolutionary!!!!

I know I'll be junking all my rb912's and rb800's if Mikrotik doesn't come out with a 802.11ac high-power card soon. I just had to hike in the snow to a tower because ANOTHER rb912 "all of a sudden" started getting ethernet errors, like the last 5 I had to replace. It's almost clockwork, like 5-6 months before they start getting real bad. Go and check your PtP's you use rb912's at, and look the ethernet stats for "RX Drops", almost every one of mine has thousands, or millions!! I can't be the only one, this is crazy now!!! And annoying...

I'd gladly keep my rb800's though and try 802.11ac, IF a Mikrotik-compatible card ever comes out! HURRY UP MIKROTIK!!!!
 
Tbird
Frequent Visitor
Frequent Visitor
Posts: 55
Joined: Tue Dec 09, 2008 8:32 pm

Re: 802.11ac

Sat Feb 15, 2014 11:26 am

We have been running AC for 7 months now, with superb stability and performance. Don't know about UBNT performance for AC, but we see even better performance than what you indicate: 150 Mbps at 20MHz, 300 Mbps at 40MHz and 620 Mbps at 80MHz. Tested with both RFC2544 tester and Mtik BW-test.

I know Flashnet is rolling out AC for both PTP and Base Stations, with excellent results, and they claimed 180 Mbps full duplex TCP. But, none of this is Mikrotik - so time is of essense to beat or match UBNT to the market for AC.

/TB
 
ste
Forum Guru
Forum Guru
Topic Author
Posts: 1811
Joined: Sun Feb 13, 2005 11:21 pm

Re: 802.11ac

Sat Feb 15, 2014 4:01 pm

We have been running AC for 7 months now, with superb stability and performance. Don't know about UBNT performance for AC, but we see even better performance than what you indicate: 150 Mbps at 20MHz, 300 Mbps at 40MHz and 620 Mbps at 80MHz. Tested with both RFC2544 tester and Mtik BW-test.

I know Flashnet is rolling out AC for both PTP and Base Stations, with excellent results, and they claimed 180 Mbps full duplex TCP. But, none of this is Mikrotik - so time is of essense to beat or match UBNT to the market for AC.

/TB
You get this speeds in lab or in real life?
 
wispwest
Member
Member
Posts: 477
Joined: Tue May 19, 2009 3:48 am

Re: 802.11ac

Sat Feb 15, 2014 7:26 pm

We have been running AC for 7 months now, with superb stability and performance. Don't know about UBNT performance for AC, but we see even better performance than what you indicate: 150 Mbps at 20MHz, 300 Mbps at 40MHz and 620 Mbps at 80MHz. Tested with both RFC2544 tester and Mtik BW-test.

I know Flashnet is rolling out AC for both PTP and Base Stations, with excellent results, and they claimed 180 Mbps full duplex TCP. But, none of this is Mikrotik - so time is of essense to beat or match UBNT to the market for AC.

/TB
Can I ask what equipment your using?? Is it outdoor gear? Is it a PCI card with MIkrotik??? Anything I can buy and try or is it beta stuff????
 
Tbird
Frequent Visitor
Frequent Visitor
Posts: 55
Joined: Tue Dec 09, 2008 8:32 pm

Re: 802.11ac

Sat Feb 15, 2014 11:01 pm

Tested in lab with RFC2544 tester and in "real life" - though the lab guys also consider themselves living in real life 8)
Like I've pointed out in AC-discussions before, good antenna spec is critical - we use 25dBi with integrated alu enclosure - and better than -45dB port to port isolation.

This is Mikrotik forum, so promoting brand names is not allowed.

/TB
 
Lakis
Forum Veteran
Forum Veteran
Posts: 701
Joined: Wed Sep 23, 2009 7:52 pm

Re: 802.11ac

Sun Feb 16, 2014 8:16 pm

AC-antenna???

promoting brand names is allowed. pls give us more info.
 
server8
Member
Member
Posts: 396
Joined: Fri Apr 22, 2011 1:27 pm

Re: 802.11ac

Mon Feb 17, 2014 9:57 am

AC-antenna???

promoting brand names is allowed. pls give us more info.
+ 1 :-)
 
Tbird
Frequent Visitor
Frequent Visitor
Posts: 55
Joined: Tue Dec 09, 2008 8:32 pm

Re: 802.11ac

Mon Feb 17, 2014 11:36 am

Not any AC specific "AC antenna" - didn't say that, but many users are not aware of the importance of Port to Port isolation for good MIMO performance. With AC and 256QAM, its even more important. Flashnet in Austria (WISP) have done a lot of testing on AC, and can probably provide more on hard facts and results. AC with 25 dBi integrated alu enclosure - I'm sure you can find it...

/TB
 
ste
Forum Guru
Forum Guru
Topic Author
Posts: 1811
Joined: Sun Feb 13, 2005 11:21 pm

Re: 802.11ac

Mon Feb 17, 2014 12:34 pm

Hint's are enough. Found it ;-)

Who is online

Users browsing this forum: MSN [Bot] and 8 guests