Broadcom introduces Wi-Fi 7 chipset solutions | Hacker News Reader

Broadcom introduces Wi-Fi 7 chipset solutions

linuxgizmos.com

93 pointsteleforce67 comments

Can we focus on longer range and penetration instead of higher bandwidth nobody needs at frequencies that require line of sight? What about the 900MHz band?

Personally my group is working on a new reliable wireless based on wireless polarization of the EM waves with collaboration with Hitachi at 900 Mhz. The initial results are very promising [1]. It's already proposed to the IEEE but perhaps at the time the timing is not right, and it's not been very well received [2].

Recently our group have improved on the wireless system and the bandwidth is much better with MPSK rather the original BPSK. From simulation the improved wireless system is a few order of magnitude better than LoRa in term of BER performance with significantly improved range especially in multi-path environment. We are now trying to improve further with additional channel coding and more efficient modulation for better efficiency in real world application with linear power amplifier (PA).

Really hope this new polarization based wireless technology can get into the upcoming Wi-Fi 7 and 6G networks, to significantly improve wireless communication reliability for consumer and industry.

[1] Successful prototyping of a rotating polarized wave receiver for high-quality wireless communication in manufacturing environments:

https://www.hitachi.com/New/cnews/month/2017/09/170911.html

[2] Introduction to RPW Experimental Study of RPW system:

https://mentor.ieee.org/802.11/dcn/18/11-18-0357-00-0000-exp...

Hi, I participated in IEEE 802.11 a few years, 10 years ago, there is a lot of politics and you have to approach the main actors (at the time Broadcom and qualcomm) and discuss with them in bilateral meetings for maybe a year, to understand what they would consider or not.

Normally if the are business cases, they should be interested and help you. Yet because it's mainly about politics you will find one or two strong groups opposed to your group for whatever weird reasons, but mainly because they do not know you: You are outsiders.

After yet one or two years of 802.11 discussions, you will succeed because they will accept you as insiders.

Just for the anecdote, the internal wars within 802.11 were quite nasty for 802.11n in 2007, and when the group met at Hawaii, suddenly all people agreed. One participant said that seing all those nude breasts was an important factor for the adoption of the 802.11n proposal!

802.11n could have been ratified ~2 years earlier, but was blocked by Texas Instruments with their ridiculous spread-spectrum but-not-really standard. It's a good thing they lost, pity it took 2 years. OFDM won, as demonstrated by the following standards n,ac,ax (and 4G/5G cellular standards).

The topic of development culture and processes in highly technical industries is fascinating to me. The Intel and Samsung drama is great.

HA seems to be the only place where it crops up - is there a decent subreddit or other forum with further discussion?

You should talk with the folks at https://eridan.io/product/ .

They're shipping[1] 16384 QAM and their scheme is a much more power and bandwidth efficient than a linear amplifier. (In addition to the coding efficiency, they don't need guard bands.)

[1] That page says "14 bits or more". I can't tell you how many bits "or more" they can do but it is non-zero.

The 5GHz band is the only usable band where I live because the 2.4GHz bnd has longer range and better penetration. There are at least 5 visible networks on each of the 3 non-overlapping 2.4GHz channels from any given point in my house, and if I'm in a corner of the house (my AP is centrally located) there are more.

Longer range and better penetration means even more congestion in built-up areas. 2.4GHz is as low as you’d want to go in any application other than rural mansions.

yeah the switch to 5ghz was huge when i lived in an apartment. i had an ap in each room and i barely received any congestion in that spectrum because the walls blocked it all. i imagine it’s great for schools and businesses in the same way.

I’ve lived in several plaster and lathe buildings and other houses that would've benefited from longer range WiFi

My house is solid brick walls throughout. My WiFi is awful in any room without an access point.

Brick walls don't block wifi that much - there's probably steel or concrete (or both) in your walls and floors. That kills wifi pretty good. In a traditional brick house you can get wifi coverage on 2.4 GHz, and for the most part 5 GHz as well, into the last nook and cranny with a single, well-placed AP.

There's definitely no steel in my house, however there is lime plaster which is somewhat similar to concrete.

I can get 2.4GHz fine, but it's pretty slow for a 1Gb/s connection. 5GHz only within the room of the AP in general (or the room immediately above/below, because the floors are just wood).

We in wall access points (eg Unifis) are great for room coverage, and there are Ethernet ports on them too, so they provide a convenient switch too.

However doing a whole house wouldn’t be cheap.

Unifis? Are they homeplugs? (Ethernet over Powerline)

Homeplugs are awesome. Used them in Singapore due to concret walls everywhere which made Wifi basically useless. Had them in every room for Wifi + Ethernet.

> Unifis? Are they homeplugs? (Ethernet over Powerline)

No, they are accesspoints

> Homeplugs are awesome

They CAN be awesome. Depends on the wireing. But if they work, they work well. I just had to replug mine every week or so but this weas a few years ago. I ethernet hardwire everything now as Im just in a small room.

Shouldn't we be able to trade off the higher speed for less conflict by having each client spend less time taking up air space? Like if we can achieve 5GBps, than presumably there could be dozens of networks operating at ~200MBps speeds all sharing the same frequencies.

Wi-Fi 6 and newer opted for OFDMA instead of TDMA but it achieves the same goal of splitting the bandwidth dynamically based on client workload. OFDMA isn't coordinated over multiple APs though.

Problem with low frequency is that it penetrates everything and has the longer range. In urban environments the spectrum will just become congested to the point of it being useless (this already happens with 2.4GHz in big cities)

Is this true even with beam forming? Or does beam forming not work well with lower frequencies?

If there are too many ssids present and they're configured to allow low data rates (which they often are) even just the beacons broadcasting the existence of the networks can consume a reasonable portion of available airtime.

Saying nobody needs it is like saying who needs more than 640k of ram. Use cases are already there in AR/VR, and additional will follow as the technology becomes available.

Yeah, modern video cable standards are approaching 100 gbps. There's plenty of potential use for way more networking bandwidth than consumers have current access to. There may not be current applications, but I think that's much more to do with lack of access than lack of utility.

Wi-Fi HaLow 802.11ah. 902–928 MHz in North-America, 863–868 MHz in Europe, 755–787 MHz in China.

Pretty low bandwidth though. Mainly focussed on IoT.

I'm following https://www.morsemicro.com/ who is implementing it from the ground up. Some amazing talent there.

Wifi is not a replacement for cellular networks. As more and more people have wifi access points it makes sense to focus on speed and congestion management.

What I would like to see is the ability for wifi devices to relay others without having access to the data. Similar to a mesh network that every device supports and some sort of decentralized management allowing better signal for all devices.

This is already in place for IOT networks like Thread, but it’s only practical at very low bandwidth and where the routing devices are connected to a power source.

I, for one, would not appreciate someone burning through my battery to stream Netflix…

WiFi is extremely power-hungry and most devices with WiFi are battery-operated (phones, tablets) - an arbitrary random mesh would only lead to dead batteries in a matter of hours.

I live in a four-plex apartment, but surrounded by other low-rise apartment buildings.

There are at least 27 WiFi networks in the 2.4 GHz bands that all dance to share bandwidth. 5 GHz is dramatically better for us, more channels and less contention.

For cellular networks, however, the only reliable service is Verizon, because they have lower-frequency bands that can provide more range out here.

Do you have use cases in mind?

Unfortunately, you’re not going to get long range without more interference. Which makes it hard to deploy with any sort of density.

I have two APs in my house and get zero reception in the bathroom that's next to one of the APs because of the tiled lath and plaster walls. Same with the kitchen in my grandparents house.

Companies like Unifi have only just released their Wi-Fi 6 APs (I think).

It feels like we've only just got Wi-Fi 6, and adoption rates in my home and office are currently very low.

Perfectly working AC AP's are not going to be upgraded for years!

Wrong. WiFi 7 has some cool stuff for location tracking.

From a previous post. https://staceyoniot.com/the-next-big-wi-fi-standard-is-for-s...

You will see WiFi 7 rollout heavily in business due to its benefit for real time asset tracking. Think warehouses, hospitals, education campuses, industrial plants. Location based asset tracking in real time has a great ROI in these areas to name a few. Its gonna be awesome.

> You will see WiFi 7 rollout heavily in business due to its benefit for real time asset tracking. Think warehouses, hospitals, education campuses, industrial plants.

I think you are underestimating the cheapness of those institutions. Maybe it says more about the places I have worked than the world at large.

Either way, we will see who is right before long.

And tracking consumers in stores

i thought the consumers were the assets being tracked.

This, along with facial recognition, needs to be made illegial in every state.

Disagree. Selling devices that you don't own and you cannot control (ie disalbe WiFi chip so tracking is not possible) should be the illegal part.

That's orthogonal to the issue, you don't need to make a technology illegal to make an application illegal.

Why not both?

WiFi 7 proposed bandwidth is just bonkers - 40G. I’d imagine that we’ll see new wireless applications like wireless docking, displays, etc.

Realistically, if we're docking next to a monitor at a desk do we really want to congest wireless networks when we can plug in for the best latency?

> Realistically, if we're docking next to a monitor

I sincerely hope that the future doesn't involve wires. Incremental improvement will be required to reach that future. This is one of them, so I welcome it.

I completely agree, but for things that "just need to work" I'll be using wires until there is no noticeable difference in using the two.

Wires also provide power

Wireless can also provide power. Note that i'm not saying it's a good way, as it has very low efficiency.

Wireless can provide power, but currently only at extremely short distances. At which point you need a wire to the device providing the wireless power anyway, so you may as well just plug it in.

True, but even now, I only charge my devices while I'm sleeping.

Exactly. People asking for more penetration/range are asking for higher power devices that interfere with each other. In an increasingly interconnected and "smart" world, where even doorbells and light bulbs are networked, you really do not want higher penetration. Think it was bad getting interference on your headset when you put something in the microwave? Imagine how bad it would be if everyone had a "smart" home and all of those devices on the block were competing with each other for the same spectrum.

Lower power, lower latency, lower penetration at reasonable bandwidths and lower costs should absolutely be the goal.

That's not to say high penetration wireless doesn't have it's uses, it absolutely does, but those are few and far between in a consumer usecase, especially if you can run an affordable lower power, lower penetration meshnet that doesn't cause interference with your own devices, or with your neighbors.

40G over what distance, three inches?

Within the same room at least, the big catch is that that’s with 16 spatial streams. A more likely 2 streams would top out at 5gbps link rate, and somewhere around 3gbps of real throughput.

Then the only real caveat is that you’re using 2/7 (or 2/3 in the EU) of the allocated 6ghz spectrum, but that’s workable given the range of 6ghz.

Not that new – WiGig has been used for VR (i.e. displays but head-mounted): https://www.vive.com/us/accessory/wireless-adapter/

Apples to oranges, very different characteristics. WiGig (802.11ad)'s up in the 60GHz band where even the oxygen in the air is absorptive, I think most walls pose big problems. But there's 6 channels of bandwidth each with 2160 MHz of space.

Wifi by compare recently went from 480 MHz of space (5GHz) to 1600 MHz (+6E spectrum) of space available in total. Colossal gain, nearly tripled in bandwidth. But there's still less bandwidth in all wifi than there is in a single wigig channel.

That wifi is making such excellent use of such a relatively limited amount of spectrum is amazing. I'm expecting impressive connectivity even at significantly reduced link budgets.

But as for the point here, about interesting new inter-device connectivity, I agree with both of you. Wigig did have these capabilities, has had interesting video & device shipping for a while now. Alas adoption has been just critically low. VR was basically the only thing that pulled this impressive wigig standard out of the hole it had crawled in to die. No one was building this stuff. Huge loss, huge pity. I really wish we had a lot more experience using wigig like technologies, because to the grandparent posters point, having this massive connectivity capability on tap for everyday consumer & business wifi: it's definitely going to enable intense new streaming & inter-system capabilities. I just wish we'd gotten more of a chance to explore this space more freely in the past 10 years, which is about how long 802.11ad/wigig's been around for.

Sure the concept has been around, but not widely implemented. I participated in a WiGig laptop dock PoC a few years ago - it wasn’t great. Usually throwing more bandwidth at stuff helps.

What's more bonkers is that bandwidth requires 16x16. Even through Wi-Fi 6 clients have remained 2x2 more than ever before.

I believe a new 6E standard is supposed to be rolling out as well.

6E is basically the same as 6, but with more frequencies available so more devices can operate without interfering. Could be a nice upgrade in a crowded environment, but in many cases I can see why people would skip it.

  # of non-overlapping 20 MHz channels

  2.4 GHz:      3
    5 GHz:      minimum 9
    6 GHz (EU): 24
    6 GHz (US): 59

source: https://www.youtube.com/watch?v=SrVTzRgi8uA

tl;dr 6E is an absolute game-changer for wi-fi congestion

Yup, 6 GHz

Now can we get driver support for older chipsets, please? What is the deal with taking a dump on the people that buy chipsets in order to use them to their advertised potential?

Is there even a draft of IEEE 802.11be out yet?

no, but the task group is headed by a Qualcomm rep and a Broadcom rep is a vice chair, so they probably have a pretty good idea of what's going to be in the final spec

https://www.ieee802.org/11/Reports/tgbe_update.htm

A lot of these will be used in prototype hardware, I'm sure the consumer ones will be the next generation

Is there public discussion of IEEE 802.12bf / WiFi Sensing?

Well I guess >1gbps internet is becoming more common so we’ll be needing it soon

What low level access is available on the chips?

Given it’s Broadcom, my guess is whatever you can reverse engineer.

Or sign away your first born and a meter of NDAs.

Reply on news.ycombinator.com