What is Wrong With Quantenna?

[RMerlin]

The benefit of XStream is that your 72 Mbps phone won't slow down the performance of your 1.3 Mbps desktop anymore. It's a brute force type of solution, but it works. And unlike MU-MIMO, it's working already with no client upgrade required (which would also cost money).

Sent from my Nexus 9 using Tapatalk

 

[azazel1024]

Linksys and Amped Wireless both are shipping AC2600 class 4x4 routers based on QCA chipsets that have working MU-MIMO. [EA8500 review.] There should be more coming out in the next few months.

MU-MIMO products using Broadcom chipsets will probably be late this year or early next.

To date, Quantenna has not released code that enables MU-MIMO.

You should not expect to see many MU-MIMO cards or USB adapters. MU-MIMO is primarily targeted at mobile devices (smartphones, tablets), so will be mainly embedded.

I would be a little surprised if whatever Intel releases to replace the 7265AC chipset doesn't support MU:MIMO. Though considering Intel's general pace of release, it'll probably be 18-30 months before we see a replacement for the 7265AC.

I am still kind of pissed that Intel is not releasing the 7265AC in mPCI-e format (only m.2 and embedded).

 

[azazel1024]

Thanks for providing your references.

There are MU-MIMO enabled client devices slowly rolling out with QCA devices inside. I listed them in the Linksys review. But again, your point is valid that there are not many available, and no cards or dongles at all that consumers can buy. That is going to remain the case for at least the short term.

I think you are misinformed about standard AC beamforming support. It is supported in 5 GHz in all routers I've seen.

MU-MIMO does not increase range. In fact, it needs strong signals to provide significant total bandwidth utilization improvement.

Finally, Quantenna may have MU-MIMO working. But they have not released the code to any of the manufacturers that produced first-generation 4x4 AC2350/AC2400 class routers. Those products remain MU-MIMO ready.

Well, you do have the possible boost to range (I said possible, I didn't say real world) because of the extra radio chain, potentially increasing signal gain, as well as possible signal gain from beam forming.

I thought that explicit beamforming support was mandatory in 11ac, I didn't think it was optional (it is optional in 11n and little supported it)? Or at least client support of it is mandatory. Obviously some routers do not support it (an AC600 router couldn't since it only has one radio chain). Most or all routers that are at least AC1200 do seem to support it, at least on 5GHz and I am pretty sure some are supporting it on 2.4GHz as well (I forget who, but I think Broadcom is a definite no on 2.4GHz support, but I want to say QCA may possibly support 2.4GHz EBF, but I could be wrong about that).

 

[thiggins]

I thought that explicit beamforming support was mandatory in 11ac

Beamforming itself, both single and multi-user forms are optional in 802.11ac. But if implemented, must use standard methods.

It was also optional in 802.11n, but there was no standard method mandated. This is why it was of limited use.

Remember the 2.4 GHz part of 802.11ac routers is still 802.11n, or at least a bastardized form of it with 256 QAM support. So beamforming isn't mandatory there either.

In actual implementation, vendors are using 802.11ac chipsets in AC1900 and higher classes on both bands for routers, so could support AC-standard beamforming in 2.4 GHz, IF clients support it, which limits potential support in 2.4 GHz to 11ac clients.