[ath9k-devel] Any 4x4 MIMO ath9k NICs coming soon?

[ath9k-devel] Any 4x4 MIMO ath9k NICs coming soon?

[Ben Greear]

I've some customers asking for 4x4 MIMO support for our wifi testing
gear (which requires the ath9k virtualization features).

Anyone know if Atheros is going to be offering a 4x4 MIMO solution
anytime soon?

Thanks,
Ben

-- 
Ben Greear <greearb@candelatech.com>
Candela Technologies Inc  http://www.candelatech.com

[ath9k-devel] Any 4x4 MIMO ath9k NICs coming soon?

[Galen]

As far as I know, Quantenna is the only producer of 4x4:4 chipsets. Some other companies have "4x4" chipsets that are 4x4:3 and simply use the additional antenna for beamforming, diversity, etc.

Historically, Atheros has begun chest-beating about chipsets significantly in advance to their release. I think the AR9003 platform started being talked about it in general term around 18 months prior to general market availability of AR9003 products.

I suspect you will have a very long wait if you are counting on Atheros (now Qualcomm) for 4x4 solutions. Interestingly, Qualcomm did announce 4x4:4 solutions a few years ago, but nothing came of them. They since acquired Atheros, so I suspect their internal 802.11 products were very unimpressive or even deeply flawed.

The reality is that 4x4:4 solutions really provide diminishing improvements relative to cost. Every additional spatial stream causes a more than linear increase in DSP complexity, a linear increase in radio/amplifier cost, a slightly more than linear increase in power consumption (radio/amplifier+DSP), and an increase in PCB footprint. And for all this extra hassle and cost, you get a typically less-than-linear increase in throughput.

The gains from jumping from 1x1:1 to 2x2:2 were huge, each increment above that is simply a lot less bang for your buck.

I'm not saying it'll never happen, but the deck is somewhat stacked against 4x4:4 solutions. I think 802.11ac will very quickly emerge as the preferred means of adding bandwidth, with 4x4:4 being something that continues to slowly grow and only becomes seen much in the wild when 802.11ac hits.

Note that Quantenna has already updated their 4x4:4 chipsets to support 802.11ac.

-Galen

On Jan 17, 2012, at 9:12 AM, Ben Greear wrote:

> I've some customers asking for 4x4 MIMO support for our wifi testing
> gear (which requires the ath9k virtualization features).
> 
> Anyone know if Atheros is going to be offering a 4x4 MIMO solution
> anytime soon?
> 
> Thanks,
> Ben
> 
> -- 
> Ben Greear <greearb@candelatech.com>
> Candela Technologies Inc  http://www.candelatech.com
> _______________________________________________
> ath9k-devel mailing list
> ath9k-devel at lists.ath9k.org
> https://lists.ath9k.org/mailman/listinfo/ath9k-devel

[ath9k-devel] Any 4x4 MIMO ath9k NICs coming soon?

[Adrian Chadd]

Well, are you talking about:

* 11n 4x4 MIMO?
* 11ac Multi-user MIMO (where you map stations to separate spatial streams?)
* other beamforming/diversity stuff?



Adrian

On 17 January 2012 11:51, Galen <galenz@zinkconsulting.com> wrote:
> As far as I know, Quantenna is the only producer of 4x4:4 chipsets. Some other companies have "4x4" chipsets that are 4x4:3 and simply use the additional antenna for beamforming, diversity, etc.
>
> Historically, Atheros has begun chest-beating about chipsets significantly in advance to their release. I think the AR9003 platform started being talked about it in general term around 18 months prior to general market availability of AR9003 products.
>
> I suspect you will have a very long wait if you are counting on Atheros (now Qualcomm) for 4x4 solutions. Interestingly, Qualcomm did announce 4x4:4 solutions a few years ago, but nothing came of them. They since acquired Atheros, so I suspect their internal 802.11 products were very unimpressive or even deeply flawed.
>
> The reality is that 4x4:4 solutions really provide diminishing improvements relative to cost. Every additional spatial stream causes a more than linear increase in DSP complexity, a linear increase in radio/amplifier cost, a slightly more than linear increase in power consumption (radio/amplifier+DSP), and an increase in PCB footprint. And for all this extra hassle and cost, you get a typically less-than-linear increase in throughput.
>
> The gains from jumping from 1x1:1 to 2x2:2 were huge, each increment above that is simply a lot less bang for your buck.
>
> I'm not saying it'll never happen, but the deck is somewhat stacked against 4x4:4 solutions. I think 802.11ac will very quickly emerge as the preferred means of adding bandwidth, with 4x4:4 being something that continues to slowly grow and only becomes seen much in the wild when 802.11ac hits.
>
> Note that Quantenna has already updated their 4x4:4 chipsets to support 802.11ac.
>
> -Galen
>
> On Jan 17, 2012, at 9:12 AM, Ben Greear wrote:
>
>> I've some customers asking for 4x4 MIMO support for our wifi testing
>> gear (which requires the ath9k virtualization features).
>>
>> Anyone know if Atheros is going to be offering a 4x4 MIMO solution
>> anytime soon?
>>
>> Thanks,
>> Ben
>>
>> --
>> Ben Greear <greearb@candelatech.com>
>> Candela Technologies Inc ?http://www.candelatech.com
>> _______________________________________________
>> ath9k-devel mailing list
>> ath9k-devel at lists.ath9k.org
>> https://lists.ath9k.org/mailman/listinfo/ath9k-devel
>
> _______________________________________________
> ath9k-devel mailing list
> ath9k-devel at lists.ath9k.org
> https://lists.ath9k.org/mailman/listinfo/ath9k-devel

[ath9k-devel] Any 4x4 MIMO ath9k NICs coming soon?

[Galen]

Ben's company focuses on load testing applications, so I assume he's referring to 4 spatial stream systems. 4x4:4 specifically refers to 4 transmit, 4 receive and support for 4 spatial streams. The only commercially available chipset for 802.11n is from Quantenna.

There are 4x4:3 802.11n chipsets available which use the extra radio for beamforming and diversity, but these are not 4 spatial stream. You get an improvement in terms of RF performance, but the MAC layer and throughput is essentially the same as 3x3:3 solutions like AR9380 and AR9390. Since Ben's application is primarily focused on these layers as far as i know, I assumed he was looking for 4 spatial streams.

Keep in mind, beamforming and diversity is very much possible within implementations the 802.11n specification. The only thing really novel to 802.11ac is the addition of optional downlink MU-MIMO. 

-Galen

On Jan 17, 2012, at 1:39 PM, Adrian Chadd wrote:

> Well, are you talking about:
> 
> * 11n 4x4 MIMO?
> * 11ac Multi-user MIMO (where you map stations to separate spatial streams?)
> * other beamforming/diversity stuff?
> 
> 
> 
> Adrian
> 
> On 17 January 2012 11:51, Galen <galenz@zinkconsulting.com> wrote:
>> As far as I know, Quantenna is the only producer of 4x4:4 chipsets. Some other companies have "4x4" chipsets that are 4x4:3 and simply use the additional antenna for beamforming, diversity, etc.
>> 
>> Historically, Atheros has begun chest-beating about chipsets significantly in advance to their release. I think the AR9003 platform started being talked about it in general term around 18 months prior to general market availability of AR9003 products.
>> 
>> I suspect you will have a very long wait if you are counting on Atheros (now Qualcomm) for 4x4 solutions. Interestingly, Qualcomm did announce 4x4:4 solutions a few years ago, but nothing came of them. They since acquired Atheros, so I suspect their internal 802.11 products were very unimpressive or even deeply flawed.
>> 
>> The reality is that 4x4:4 solutions really provide diminishing improvements relative to cost. Every additional spatial stream causes a more than linear increase in DSP complexity, a linear increase in radio/amplifier cost, a slightly more than linear increase in power consumption (radio/amplifier+DSP), and an increase in PCB footprint. And for all this extra hassle and cost, you get a typically less-than-linear increase in throughput.
>> 
>> The gains from jumping from 1x1:1 to 2x2:2 were huge, each increment above that is simply a lot less bang for your buck.
>> 
>> I'm not saying it'll never happen, but the deck is somewhat stacked against 4x4:4 solutions. I think 802.11ac will very quickly emerge as the preferred means of adding bandwidth, with 4x4:4 being something that continues to slowly grow and only becomes seen much in the wild when 802.11ac hits.
>> 
>> Note that Quantenna has already updated their 4x4:4 chipsets to support 802.11ac.
>> 
>> -Galen
>> 
>> On Jan 17, 2012, at 9:12 AM, Ben Greear wrote:
>> 
>>> I've some customers asking for 4x4 MIMO support for our wifi testing
>>> gear (which requires the ath9k virtualization features).
>>> 
>>> Anyone know if Atheros is going to be offering a 4x4 MIMO solution
>>> anytime soon?
>>> 
>>> Thanks,
>>> Ben
>>> 
>>> --
>>> Ben Greear <greearb@candelatech.com>
>>> Candela Technologies Inc  http://www.candelatech.com
>>> _______________________________________________
>>> ath9k-devel mailing list
>>> ath9k-devel at lists.ath9k.org
>>> https://lists.ath9k.org/mailman/listinfo/ath9k-devel
>> 
>> _______________________________________________
>> ath9k-devel mailing list
>> ath9k-devel at lists.ath9k.org
>> https://lists.ath9k.org/mailman/listinfo/ath9k-devel

[ath9k-devel] Any 4x4 MIMO ath9k NICs coming soon?

[Ben Greear]

On 01/17/2012 02:03 PM, Galen wrote:
> Ben's company focuses on load testing applications, so I assume he's referring to 4 spatial stream systems. 4x4:4 specifically refers to 4 transmit, 4 receive and support for 4 spatial streams. The only commercially available chipset for 802.11n is from Quantenna.
>
> There are 4x4:3 802.11n chipsets available which use the extra radio for beamforming and diversity, but these are not 4 spatial stream. You get an improvement in terms of RF performance, but the MAC layer and throughput is essentially the same as 3x3:3 solutions like AR9380 and AR9390. Since Ben's application is primarily focused on these layers as far as i know, I assumed he was looking for 4 spatial streams.
>
> Keep in mind, beamforming and diversity is very much possible within implementations the 802.11n specification. The only thing really novel to 802.11ac is the addition of optional downlink MU-MIMO.
>
> -Galen
>
> On Jan 17, 2012, at 1:39 PM, Adrian Chadd wrote:
>
>> Well, are you talking about:
>>
>> * 11n 4x4 MIMO?
>> * 11ac Multi-user MIMO (where you map stations to separate spatial streams?)
>> * other beamforming/diversity stuff?

I'm mostly just passing on a question that a customer asked me.  I assume
they want 4x4 MIMO with 4 spatial streams.

I'll look into 11ac as well.

I'll also take a look at the quantenna offerings, but so far, Atheros is
the only manufacturer that has the driver/hardware support for the virtualization
that we need (since way back in ath5k/madwifi days), so I sort of assumed
that isn't going to change anytime soon...

Thanks,
Ben

-- 
Ben Greear <greearb@candelatech.com>
Candela Technologies Inc  http://www.candelatech.com