802.11ac Specification Is Final - so what's the future hold?

[vnangia]

Hi Tim, on a practical level, does this mean anything for us as end users?

I've been holding off on an ac router for several reasons, including a lack of clients, and more importantly, the fact that vendors seem to be (very slowly) rolling out more and more streams. Does finalization make it more likely we're going to see vendors rolling out full MU-MIMO setups, which could conceivably give us the theoretical 6.7Gb/s that we were expecting to see?

Somewhat unrelated: I was expecting to see 802.11ac/ad products roll out this CES, so high speed in the same room with fall back to lower speeds as you move out of the 60GHz band's reach. In fact, I can't think of any ad devices at all - is it DOA in your opinion?

 

[stevech]

11ad.. next try at wireless USB.
Cable replacement.
IMO: it's a long haul to get home theater/HDTV manufacturers to replace HDMI audio/video with wired/wireless something that actually interoperates/works well to meet users' expectations. On that same racetrack is 4K video (Sony) which likely exceeds what can be provided by any near future wireless.

 

[sfx2000]

Here's the 802.11 working group timeline...

http://grouper.ieee.org/groups/802/11/Reports/802.11_Timelines.htm

Lots of interesting stuff on the docket.

 

[vnangia]

11ad.. next try at wireless USB.
Cable replacement.
IMO: it's a long haul to get home theater/HDTV manufacturers to replace HDMI audio/video with wired/wireless something that actually interoperates/works well to meet users' expectations. On that same racetrack is 4K video (Sony) which likely exceeds what can be provided by any near future wireless.

I thought that's WISA and it's been standardized? Okay, so at least I now know not to expect ad to be part of a router, but a point-to-point communication replacement and not to expect routers with ad... I'm not sure how I ended up with that mistaken belief.

Here's the 802.11 working group timeline...

http://grouper.ieee.org/groups/802/11/Reports/802.11_Timelines.htm

Lots of interesting stuff on the docket.

Yeah, it really is. I hadn't realized that it was now public.

At any rate, I guess the thing that I'm wondering most of all is: how come we're not seeing a over-the-top 8 antenna, 160MHz-wide, 6.7Gb/s-capable router with 10G-baseT yet? Surely someone wants to be first in this!

 

[sfx2000]

The 802P workings groups have always been "public" if not well hidden - anybody can attend a meeting (paying the $$$) and comment, the floor is always open - attend enough of them and you get voting rights...

Near term, let's say next 12-24 months:

* cleanup of 802.11 and merger of the 11ac/ad amendments into the 802.11 baseline
* backporting of some of the 11ac features back into the 2.4GHz space (along with clearing out the non-used parts of 11n)
* better handover support within the 11 space (getting better, but room for improvement)
* enhancements for auth to replace a somewhat broken WPS/11i user experience for consumers
* formal integration of Hotspot 2.0 (more authentication improvements).

But who knows - bodies like 802P, 3GPP, 3GPP2, and IETF are the engineering versions of Congress (or some other legislative group) - everything is a compromise - to whit 802.11n which combined several initiatives into one specification... it's slow going sometimes, and very political.

 

[sfx2000]

How come we're not seeing a over-the-top 8 antenna, 160MHz-wide, 6.7Gb/s-capable router with 10G-baseT yet? Surely someone wants to be first in this!

I can build you one now - how much do you want to spend - it'll be a bit more than $200USD, which seems to be the current going rate/pain point for high end consumer gear

 

[thiggins]

First, today's announcement was just a check box that had to be done.
No one developing 11ac products was waiting for this announcement. All it means is that anyone who was still saying "draft" 802.11ac can now stop. Hell, I even stopped about a month or so ago.

No sign of Wi-Gig products at the show and no one is talking about them either. I think 60 GHz had a market window and they missed it. Why put another RF chain and antenna in a product to provide a short range connection that requires the same added circuitry on the other end?

Wi-Fi has won the mindshare contest and 11ac is increasing throughput to handle the applications that Wi-Gig had a shot at.

Cable replacement is a niche market for folks with lots of money to burn. For them, Gefen and whole home control guys have plenty of niche solutions to offer. Making a hole in the wall is still the cheapest way to hide the HDMI cable running between the STB and display. And you gotta get a power cord to the display anyway.

Now why are you waiting for more streams? What do you have that is even using the three streams now available? And you really want to spend $300 bucks on a router ($300 is the new $200 for the I-just-want-the-most-expensive-router-I-can-buy crowd).

MU-MIMO is hella difficult and, to quote Matt Gast in his 802.11ac Survival Guide "has yet to be proven in commercially available products in widespread use". In other words, it's a science experiment and all the $300 router buyers are the lab rats.

MU-MIMO depends on beamforming and a significantly more complex sounding process than the basic beamforming in standard 11ac routers. It works downstream only depends on clients to support it. (And no, there is no "implicit" MU-MIMO).

I think I'll stop the rant now.

 

[sfx2000]

11ac - MU-MIMO - it's going to be interesting with how chipset vendors manage the scheduling and beamforming - having been there in 3GPP2 and with SDMA on i-Burst (and discussion for 802.16m and 802.20) - this stuff is very non-trivial... as they say on facebook - it's complicated... and complication is cost...

11ad - think wireless HMDI and Docking Stations - back when I was working standards, had an interesting conversation with a gentleman from Dell - it was all about near-field, and this was back in '08 for millimeter wave - and most of the SME's were about the local room distance for 60Ghz - thing is at 60Ghz, what is 160MHz in a channel, not much... and the band there is pretty wide... but ranges where in the centimeters... not whole house/office... 60Ghz is bouncing off paper... and the air that we breath (that's why it's another one of the "garbage bands" for unlicensed operation, BTW)

R&D is about 5 years out from commercial deployment - just saying... nice to see some of this work hit the market...

 

[L&LD]

Now why are you waiting for more streams? What do you have that is even using the three streams now available? And you really want to spend $300 bucks on a router ($300 is the new $200 for the I-just-want-the-most-expensive-router-I-can-buy crowd).

MU-MIMO is hella difficult and, to quote Matt Gast in his 802.11ac Survival Guide "has yet to be proven in commercially available products in widespread use". In other words, it's a science experiment and all the $300 router buyers are the lab rats.

MU-MIMO depends on beamforming and a significantly more complex sounding process than the basic beamforming in standard 11ac routers. It works downstream only depends on clients to support it. (And no, there is no "implicit" MU-MIMO).

Tim,
I want more router streams not to connect to single 3+ stream clients but rather to transmit to multiple clients (even if it is just for downloading) at once. As an internet article explains it: MU-MIMO is like going from a hub to a switch, wirelessly.

I want my customers to have as few routers as possible and still be able to service all wireless requests in a timely manner (whether it's a single user or 500 devices at once). This is especially important with limited channels to chose from - even with the greatly expanded 5GHz and AC ranges (and especially if you need multiple radios in a relatively small space like a large conference room, for example.

Of course, I am discounting the $300 price now being waved around (with no product available, of course). But cost alone does not dictate whether something is worth it or not.

With these products shipping (along with your excellent reviews) we will be in a much better position to know if the price asked is worth the product as delivered. Even if it stays at the high range it may still be the better option for a dense client situation.

Sure, some of us may have to be the lab rats for this hub to switch jump in performance experiment. But I think it's a worthwhile thing to pursue.

Even if MU-MIMO doesn't work out as expected (though I have a hunch it will) - it has to be at least tried. Which will still bring us to something that will give us a hub to switch performance jump in our wireless networks, even if it isn't called MU-MIMO then.


(Thank you for the rant and the reality checks to my rose colored wireless world).

 

[vnangia]

A relative has an in-law suite and I was suggesting using a pair of 11ac routers over about 40-50ft of open air as an alternative to digging a trench and running a cable. Huge lot, very little interference, hence I figured maximize the bandwidth, and no matter what the cost of 8x8 routers, it'll still be cheaper than getting the permits to dig said trench, or running a proprietary microwave link. To be sure, he could just swing a cable, but he lives in a part of the world that sees -30C to +55C weather and violent storms to boot. Seemed reasonable to me to suggest $900 on two routers to serve as a site-to-site link.

At any rate, I'll be waiting to see about the Asus N87U availability and results. I suspect I'll suggest that and just be done with it.

 

[thiggins]

A relative has an in-law suite and I was suggesting using a pair of 11ac routers over about 40-50ft of open air as an alternative to digging a trench and running a cable. Huge lot, very little interference, hence I figured maximize the bandwidth, and no matter what the cost of 8x8 routers, it'll still be cheaper than getting the permits to dig said trench, or running a proprietary microwave link. To be sure, he could just swing a cable, but he lives in a part of the world that sees -30C to +55C weather and violent storms to boot. Seemed reasonable to me to suggest $900 on two routers to serve as a site-to-site link.

At any rate, I'll be waiting to see about the Asus N87U availability and results. I suspect I'll suggest that and just be done with it.

Quantenna based routers are at least six months away. From chatter during the show, everyone knows MU-MIMO is non-trivial and that Quantenna's experience to date has been on dedicated matched bridge type products. Very different to make a router that has to deal with the whole range of clients, not just a Quantenna based STA.

 

[Razor512]

One issue that I have noticed is that in countries like the US, we effectively only have 5 channels in the 5GHz band

the select few lower channels are heavily restricted, making them not very useful.

If we enter an environment which multiple wifi networks near by, then we will never be able to really use 802.11ac unless they find some new modulation which increases throughput without using additional channels.

AC1300 already takes up 4 of the 5 useful channels, for example on my R7000

149 + 153(P) + 157 + 161

Unless the FCC stops being jerks, we will never make it much higher than the 1300

 

[Kris404]

A relative has an in-law suite and I was suggesting using a pair of 11ac routers over about 40-50ft of open air as an alternative to digging a trench and running a cable. Huge lot, very little interference, hence I figured maximize the bandwidth, and no matter what the cost of 8x8 routers, it'll still be cheaper than getting the permits to dig said trench, or running a proprietary microwave link. To be sure, he could just swing a cable, but he lives in a part of the world that sees -30C to +55C weather and violent storms to boot. Seemed reasonable to me to suggest $900 on two routers to serve as a site-to-site link.

At any rate, I'll be waiting to see about the Asus N87U availability and results. I suspect I'll suggest that and just be done with it.

IMO, laying a CAT 5e/6 cable in such a scenario would be worth it if it's logistically feasible.

 

[thiggins]

One issue that I have noticed is that in countries like the US, we effectively only have 5 channels in the 5GHz band

the select few lower channels are heavily restricted, making them not very useful.

If we enter an environment which multiple wifi networks near by, then we will never be able to really use 802.11ac unless they find some new modulation which increases throughput without using additional channels.

AC1300 already takes up 4 of the 5 useful channels, for example on my R7000

149 + 153(P) + 157 + 161

Unless the FCC stops being jerks, we will never make it much higher than the 1300

The good/bad news is that 5 GHz range is more limited. So less chance of neighboring network interference. But this problem was pointed out back in 2012. Why 802.11ac Will Kill The 5 GHz Wi-Fi Band