Wi-Fi 6(802.11ax)テクノロジー
最新のWi-Fi規格はWi-Fi 6(技術的にはIEEE 802.11axとも呼ばれます)であり、無線通信の新時代を切り開きます。Wi-Fi 6の焦点は、最大スループットの向上のみではなく、効率と容量を最適化することです。Wi-Fiテクノロジーの未来として勢いを増しています。ネットワークにWi-Fi 6を検討すべき理由を簡単に確認するには、次のビデオをご覧ください。
Hello everyone and thank you for joining the webinar. Today we're going to talk about 802.11ax which is the newest generation of Wi-Fi. So I want to spend a little bit of time going through an overview and some of the benefits including I'll spend some time on OFDMA, show you a little bit what it looks like, some of the benefits we've seen from it, what kind of considerations you should take into account for when you design a network for 802.11ax. Is there anything different? I'll give you the spoiler right now, really there's nothing different that you need to worry about, just design like you would for 11ac.
And things that you might want to watch out for, you know, what are the timelines, is there anything with regards to 11ax that you should really be cognizant of. So what's, you know, what is all the buzz about 11ax? And really there's a couple of what we call big ticket items around 11ax that are making people excited. Probably the biggest one is OFDMA or orthogonal frequency division multiple access.
In 11ac we had multi-user MIMO which was a multi-user technique to be able to talk to multiple clients at the same time. 11ax is adding another multi-user enhancement and I'll talk a little bit about the differences between multi-user MIMO and OFDMA. Another one that we're excited about is BSS coloring, also known as spatial reuse.
And spatial reuse is a way for high density Wi-Fi networks to operate just a little bit better. Target wait time is another one, this is more about battery efficiency on battery-powered devices. The next one is uplink multi-user MIMO.
I talked about how 11ac had multi-user MIMO but this was downlink only multi-user MIMO. 11ax introduces the multi-user MIMO in the uplink direction as well as 1024 QAM. This is an increased modulation technique over 11ac and it provides just a little bit better throughput than 11ac.
So as far as timelines go, every vendor right now is talking about 11ax and it's the time to buy because everybody has new APs so they want to sell them. And really it's up to you when you want to buy but just as far as an industry goes, we're still at the very beginning of the 11ax adoption curve. We do not yet have a Wi-Fi alliance, Wi-Fi 6 certification.
That's expected to begin in Q3 this quarter this year. The Wi-Fi alliance has already announced it so you can check out some of the details on the website but there have been no products yet and the certification process has not yet begun for Wi-Fi 6. The IEEE, the governing body that actually controls the 802.11 standard has not yet ratified 802.11ax. So this is, we're still in actually draft phase. Now most of the technical details have already been worked out so we're not expecting really interop issues, just we do not actually yet have a fully ratified amendment.
As far as client devices go, I think that there are exactly two devices out there, maybe three. It depends on what you're looking at. The specs are a little sketchy sometimes but really there's two devices that we know that you can buy and so the device ecosystem is very, very low.
We're expecting that to increase over the next six to twelve months but right now it's only two devices. Then as far as APs go, APs have been coming out from vendors for the past few months, even past year but these APs that have come out haven't had Wi-Fi 6 turned on or 11ax turned on. It's only now pretty recently or if the vendors have done it at all that AX has been turned on and we're expecting more and more APs to come out again over the next 6, 12, 18 months.
So we're still at the very beginning of the life cycle and so don't feel rushed, don't feel pressured. Upgrading to AX is, do it when it feels right for your organization. So I've talked a little bit, I've mentioned 11ax, I've mentioned Wi-Fi 6 and just to talk about the terminology to clear up any sort of confusion there might be when you should refer to 11ax, when you should refer to Wi-Fi 6. Technically, the terms are interchangeable but if you want to be the most technical, Wi-Fi 6 refers to the Wi-Fi Alliance certification.
So a vendor would go to the Wi-Fi Alliance, the Wi-Fi Alliance would certify their device for 802.11ax and that device would be called Wi-Fi 6. The Wi-Fi Alliance recently changed their naming convention from matching what the IEEE did to a more marketing friendly name. So 11ax is known as Wi-Fi 6, 11ac is known as Wi-Fi 5 and 11n is known as Wi-Fi 4. And 11ax obviously is what the IEEE working group is called. So I want you to also notice what the fi names are.
So 11n was known as high throughput, 11ac was known as very high throughput and 11ax is known as high efficiency. And that's really the theme. Every previous generation of Wi-Fi has aimed to increase the maximum throughput of Wi-Fi.
So 802.11 to 11b was increased from 2 megabit to 11 megabit, 11b to 11a or g went up to 54 megabits per second, 11n went up to 600 megabits a second. Now we never saw that with 11n really, we only saw three spatial stream APs up to 450 megabits per second. 11ac technically allowed up to 6.9 gigabits per second. Again, we never saw that kind of data rate. 2.3 was really the best area that we kind of saw. And then 11ax, maximum data rate is 9.6 gigs per second.
But again, we're not going to see that type of data rate on both client devices and APs. And so the way that previous generations of Wi-Fi had increased throughput was to use MIMO, multiple input, multiple output, the ability to use multiple transmitters at one time. And so we have two spatial stream, three spatial stream, four spatial stream access points, increasing the modulation rate to going from 64 QAM to 256 QAM to now 1024 QAM with 11ax.
Also increasing the channel width. So these are all techniques we've we've gotten pretty far along the way of getting all the spectral efficiency that we can at one time. And so now comes 11ax.
And 11ax is really about increasing the efficiency of Wi-Fi. And the way it does that is through multi-user access techniques. So OFDMA, being able to talk to multiple devices at the same time. And yes, we also do have maximum throughput type features as well, such as 1024 QAM. Some of the subcarrier spacings change that actually increases the data rate. But by and large, the whole goal of 802.11ax is efficiency.
The fine name is literally called high efficiency. And so that's what we're trying to do here, is trying to increase the efficiency, decrease some of the overhead associated with Wi-Fi. And the reason is because Wi-Fi has a scale problem.
The best throughput that you'll ever see in Wi-Fi is when you have just one single client connected to an access point. And if you're using, you know, 80 megahertz channels or even 160 megahertz channels, you can see pretty good throughput, right? At a, you know, a laptop, you might be able to see 800 megabits per second with one client. But once you add that second client, and once you add that, you know, that that 10th client, that 25th client, the 100th client, the throughput really goes down pretty quickly.
And so, you know, the way that you get around this is you add more APs, you add more cells, but there's only so much, only so many channels, only so much frequency. And this is what, you know, this is what 11ax is trying to solve. It's trying to increase the per client throughput within Wi-Fi. So if we take a look at, you know, Wi-Fi today, it's very much of an unscheduled mess. It's everyone, you know, trying to contend for the medium at one time. And you run into a lot of traffic jams, right? There's a lot of efficiency.
Wi-Fi is CSMA collision detection, right? CD. I'm sorry. Yeah, collision detection, where Ethernet is collision avoidance, right? And so, sorry, I have the other way around.
Wi-Fi is collision avoidance. So it's actively trying to avoid collisions, which means that if a Wi-Fi transmitter hears somebody else transmitting, it's going to back off. And then it's going to, you know, randomly, there's a random interval.
And then if you hear somebody else transmitting it again, it's going to back off even further, right? So we get these exponential delays, which causes these big, you know, these big congestions, big traffic jams within Wi-Fi. OFDMA is giving a little bit more control over the channel to the access point. So the access point, to some extent, will be able to actually schedule what goes on in the channel.
It's going to say, I'm going to talk to, you know, these four clients, these eight clients at the same time, or conversely, it can say, okay, you four clients, you eight clients, you're able to transmit to me right now. And so it adds a little bit of, you know, order to an otherwise chaotic technology. And we're expecting, you know, big efficiency gains because of that.
So again, here's the big ticket items within Wi-Fi. I'm going to go through some of these in a little bit more detail. So what is OFDMA? So Wi-Fi today, typically most data rates are sent using what's called OFDM modulation, right? Or orthogonal frequency multiple division, orthogonal frequency division multiplexing.
And so if you think about it, if just like a two-way radio, if only one person can talk on the channel at a time. So if I have three clients to talk to, imagine each one of those clients, I need to transmit separately to each of those clients. And so the example here is I need to send three trucks down the freeway to talk to those three clients.
With OFDMA, what I'm doing, right, OFDMA is multiple access. I'm literally splitting the channel into smaller chunks, right? So I'm splitting the channel into smaller, what's called resource units. And so I can give half the channel to one client if it's, you know, downloading a video, another client might be doing a voice call.
So I'll give them, you know, a quarter of the channel and then another client's surfing Twitter. And so I'll give them the other, the last quarter of the channel. And essentially, so I can talk to those three devices at one time in one transport opportunity, right? I only need to send one truck down the freeway and I greatly reduce my overhead, right? I don't need an extra driver.
I don't need an extra tractor for this, you know, to send that truck. It greatly reduces the overhead within Wi-Fi and it allows, you know, better efficiency, more data to be transmitted because we're making better use of the airtime. So that's OFDMA.
Now what's the difference between OFDMA and multi-user MIMO? So think of OFDMA as frequency multiplexing, right? So multiple users talking on the channel at one time by splitting the channel into smaller chunks. Multi-user MIMO again is a multi-user technique, but instead of splitting the channel into smaller chunks, I'm actually splitting my spatial streams. And so access points have multiple spatial streams.
Most high-end APs today are four spatial streams. So I can give one spatial stream to one client, one spatial stream to another client, and my third spatial stream to another client. And this is great.
It's a, you know, in theory it works. We haven't actually seen it pan out so much in the real world. There's been some clients haven't supported it and it seems kind of finicky and you have to have, you know, all the planets lined up to get it to work actually properly in the real world.
But fundamentally this is the difference, right? OFDMA is splitting the channel into smaller chunks. Multi-user MIMO is dividing the spatial stream among the clients. So I mentioned the term resource units and this is kind of an important point to understand regarding OFDMA.
So the OFDMA breaks up the channel into smaller chunks and it doesn't do it randomly. There's a kind of, there's a, there's a formula to follow and this is the way that a channel can be broken up into smaller chunks. So a, the channel can be broken up into, you know, 106 resource units.
So essentially you've given the client half the channel, right? So, you know, what is it, about eight megahertz. I can, I can break the channel up into quarters and essentially I've given the, I've given the client four megahertz out of the 20 megahertz channel. Or I could break it up into eighths or ninths and, you know, essentially give up two megahertz of the channel.
And again, this is for 20 megahertz. For 40 megahertz, same kind of, same kind of concept of the, you know, 102, 50 and 26 resource units, except I can have up to 18 users in a 40 megahertz channel. And then an 80 megahertz channel, again, we still get down to two megahertz, 26 resource units is the smallest that we can go.
But because the channel is so wide, I can have more users. And that's, you know, that's OFDMA in a nutshell. And so let me, you know, we've, we've done some testing with OFDMA.
We've actually seen it work with the clients out there. Pretty cool to see. So if you're familiar with Wi-Fi, you're likely familiar with what an OFDM waveform looks like, right? This is, this is an OFDM waveform, you know, pretty flat on the sides.
It goes up vertically and then you have a flat, you know, flat horizontal where, you know, across the entire channel width. So this is a 20 megahertz wide channel, pretty standard. And this is what OFDMA looks like.
So do you notice the difference here? You know, the sides really look the same. The difference is we have this null in the middle, right? And that's because if I go back here to my 20 megahertz, right, we have these, you know, these neutral carriers in the middle of the channel. And so that's, that's actually what we're seeing here.
This is with two clients. Each client has 106 resource units and, you know, is essentially been allocated half the channel. And so that really the dead giveaway that you can spot OFDMA is this null in the middle of the channel.
And so kind of real, real world testing, we've seen OFDMA, it actually works fairly consistently. It's not 100% of the time of the transmission. You know, it seems to, it seems to go in and out, and that'll probably improve as drivers improve, but we're actually seeing it in the real world with the clients that exist.
And you don't have to do anything special to make it work, which I think is really, really cool. And here's, if you want to spot OFDMA by taking a PCAP, it's, it's really difficult. You, you likely will not actually be able to capture the OFDMA data frames, but you can capture the OFDMA management frames.
And so here is a trigger frame. And if you want to look this up, you know, in Wireshark, if you have clients in a 11 AXAP, here's the frame control, it's a hex value of 2400, and you'll be able to see the trigger frames. And so in this case, I have four OFDMA users, and the AP, this is the AP actually saying, okay, these four users, I'm going to expect OFDMA, and here's the resource units that you've been allocated. In this case, 52 resource units. So that's OFDMA. Let me talk a little bit about BSS coloring.
BSS coloring is, overcomes Wi-Fi's inherent politeness. So Wi-Fi is a, is very much of a listen before talk technology, which means that if there is a neighboring signal in Wi-Fi, you know, somebody else on the channel, that AP or client is going to back off, right? It doesn't want to transmit over those devices, even though it could potentially transmit over those devices without causing any sort of issue. Think of it as a, you know, you're, you're, you're in a noisy bar, right? There's a lot of other people having conversations.
And the way Wi-Fi works today is only one person in that bar could be able to talk at a time. So when you have a lot of people in one environment, you can tell that that would, you know, that could cause issues. You know, Wi-Fi won't transmit very frequently.
And so to overcome that, you know, Wi-Fi is essentially implementing a technique that you could have a conversation with your neighbor and just ignore everybody else in the bar. And this is called spatial reuse or, or VSS coloring. And the, really the underlying technique is that clients will maintain, and APs will maintain a second nav timer, right? So a nav timer is, is what the, what the station or AP uses to back off or get off, you know, for, to, to, to get on, onto the channel.
And so that second nav timer, there's a concept of a overlapping VSS and my VSS. And essentially what it's doing is adjusting the CCA threshold for when a client says, you know what, this is a overlapping VSS. I don't really care about this.
My nav, my CCA timer is way more aggressive. I'm going to get on the channel because it doesn't matter if I talk over these guys, but this is somebody that has, you know, my same color. I should, you know, I should back off, you know, it would actually be detrimental if I transmitted over these guys.
And so this is kind of just a visualization. So notice that we have two APs on channel one, and they both have the gray VSS color. And so in this, in this instance, two APs that have the same VSS color would follow the normal channel access rules.
Then I have, you know, two APs on channel 11, but notice they have different VSS colors. So in this case, the APs would have, you know, utilize that second nav and be able to, you know, get onto the channel, get onto the channel a little bit more aggressively to overcome that, that, that politeness that I was talking about. 1024 QAM.
So 1024 QAM is, is an enhancement to the modulation techniques within Wi-Fi. And in 11AC, we had 256 QAM, which was an enhancement over 64 QAM in 11N. So we keep increasing the modulation. We keep putting more bits onto the air at the same time. And so essentially 256 QAM is eight bits per symbol. 1024 QAM is 10 bits per symbol.
So 25% improvement in terms of bits on the air at the same time. The downside is you need a really strong, a more strong signal strength in SNR to be able to decode 1024 QAM. And so it varies on the AP and the data rates and the channel width that's in use.
But typically 1024 QAM needs about neg 60 dBm to decode where 256 QAM needs about neg 68 dBm. And so this map, the colored area is an estimated 1024 coverage area versus the gray area, which is an estimated 256 QAM coverage area. And so 1024 QAM is great if you can actually transmit it and have the clients decode it.
And the range that we've seen is fairly small, typically within 15, maybe 20 feet, if you're lucky, where 256 QAM could be anywhere. You could potentially maintain that up to 30, 40 feet. So there's a bit of a difference there.
And don't expect 1024 QAM everywhere within your cell. And certainly don't design for 1024 QAM because you would put in way too many APs and cause too much co-channel interference and your clients wouldn't want to roam very well. As far as data rates go, the 11AX is actually using a completely different layer one transmission.
So 11AX HE data rates are completely different from 11AC VHD data rates. So if you looked at 11N versus 11AC, the data rates, 11N data rates were the exact same as 11AC. MCS5 for 11N was the same data rate as MCS5 for 11AC.
Only difference is 11AC added MCS8 and 9 with 256 QAM modulation. With 11AX or high efficiency, the data rates are completely different across the board. And that's because there's different sub-carrier spacing.
So 11N and 11AC used a sub-carrier spacing of 312.5 kilohertz. 11N, or sorry, 11AX uses a much narrower channel sub-carrier spacing of 78.128 kilohertz. As well as there's different guard intervals, right? So 11N and 11AC, you could do 400 and 800 nanosecond guard interval.
11AX, you can do 800, 1600, or 3200 nanosecond guard interval. And so all those changes, that's about a 20% improvement in terms of data rate for 11AX data rates. So the chart here shows you just a little bit of the differences, right? So MCS0, 11AC, that data rate would be 7.2 megabits per second.
11AX MCS0 is 8.6 megabits per second. So you should expect to actually see a little bit of a throughput bump because of the increase in data rate with 11AX. So that's kind of the technology overview.
I want to talk a little bit about some of the trends and what you'll see out there from vendors. And so we've already started to see it, this concept of four by four access points versus eight by eight access points. 11AX, just like 11AC, allows up to eight spatial streams.
So that's a eight by eight, colon eight, that's the notation that you would see. So these are really big access points, highly powerful access points. The problem is the first generation of eight by eight APs, which happen to be Qualcomm-based, do not support uplink OFDMA.
They cannot be Wi-Fi 6 certified because they don't support uplink OFDMA. And so that's kind of a ding in first place against eight by eight. Now there is a second generation of Qualcomm APs out there that does support uplink OFDMA or will be enabled at some point.
But so as a customer, you kind of need to decide, do I care about eight by eight? What are the potential benefits? Potential benefits, three areas. You have potential for better RF performance due to beam forming, as well as in the uplink direction MRC. Four by four is already pretty good in this regard, but you would actually get a little bit of a benefit over that with eight by eight.
If you had eight by eight clients, it would make sense to have an eight by eight AP. It's very unlikely that we're going to see even one eight by eight client, let alone multiple. We're expecting eight by eight clients to be boutique type clients if they exist at all.
And what we're seeing is with today's, today we have four by four APs and four spatial stream clients. And it's actually very difficult to maintain four spatial streams with a four by four AP. We're expecting it to be even more difficult to maintain eight spatial streams with an eight by eight AP.
So it's just some things to keep in mind. You hear the term eight by eight, but there's no guarantee that there's going to be eight by eight clients or that you'll be able to maintain eight spatial streams. As well as multi-user MIMO.
Multi-user MIMO, most clients are two spatial stream. So you could have three or four clients in a multi-user MIMO group and potentially see quite a good benefit. The problem is thus far multi-user MIMO has proven to be ineffective.
It's not worked very well in the real world. At least it's not proven. There are devices out there, but we're just not seeing multi-user MIMO in the wild. It's just not happening. And so really you can negate those bottom two options. And the first benefit is the only benefit of eight by eight.
So from my perspective, eight by eight is kind of a, let's wait and see how it pans out kind of move. I certainly wouldn't buy an AP today that does not support Uplink OFDMA. It's a major feature of AX, a major reason to buy AX.
So just whatever vendor you go with, just ask if your AP supports or will support Uplink OFDMA. Eight by eight APs will be likely to require more power. Some of them will require BT power, not all of them.
Some of them will operate under AT. So just another thing to keep in mind if you're going to need to upgrade your switches as well. And lastly, an interesting dynamic that I'm excited to see play out.
I have my own opinions, but there are APs out there, MIST is one of them, that has dual five, four by four, right? So dual five gigahertz capable AP. So dual five, four by four versus an AP that has a single eight by eight, five gigahertz AP or radio. And I'm really excited to see how that's going to play out.
My kind of thought is dual five, four by four will be more beneficial in the real world than a single eight by eight, because essentially you have two cells out of the AP, you're going to have increased capacity. But we'll see how that plays out over the next six or 12 months. Do you need to upgrade your switch for 802911 AX? Three reasons to upgrade a switch.
The first is if your existing switching infrastructure is end of life, you may want to think about upgrading due to obvious reasons, security patches, whatnot. But typically what we see from customers is realistically they'll need to upgrade their switches due to PoE. And this is probably the number one issue.
If you have AF switches, most likely you'll need to upgrade to AT for 11AX. Most of the APs out there will operate under AT power. Some might need BT, including eight by eight APs when you enable the USB functionality, or if the AP supports a PoE out functionality.
Then secondly is multi-gigabit. And this has been around, you know, this conversation, you know, we also had with 11AC. And my opinion is reality is a customer will never exceed one gig per second in the real world. So my thought is one gigabit per second is sufficient. 2.5 gig will be the norm. Most APs will have a 2.5 gig port and some APs might even have a 5 gig port.
But reality is you'll be fine with a gig port. There might be, you know, some corner cases where you could exceed a gig, but very, very rare in the real world. And here's just some math behind it, right? So a 20 megahertz channel, you will not exceed a one gig per second, not even with four spatial streams, dual five or eight by eight.
40 megahertz channel, if you're eight spatial stream, M7, AM8 could exceed a gig, four stream dual five will not exceed a gig. But again, that assumes that you have an eight spatial stream client, which does not exist. And then 80 megahertz channel, you could exceed a gig threshold at three spatial streams, M10 or M11.
Just to wrap up, you know, this point, it is possible to exceed a gig in corner cases in the lab. We're not expecting it, that to be the norm in the real world. So frankly, I wouldn't worry about, you know, that gigabit per second Ethernet port.
If you're worried about future proofing, whatever it may be, then it makes sense to upgrade. If you want the very best, it makes sense to upgrade to M gig, but you don't need it by default. And how about 160 megahertz channels? This was supported in 11AC, also supported in 11AX. Just don't do it. That's the takeaway. If you're in your house, maybe.
But in the enterprise, in business, we have 25, 20 megahertz channels in the U.S. That's a lot of channels. You have a lot of capacity with that amount of channels. You can even use 40 megahertz channels, right? We dropped down to 12 non-overlapping channels because we lose channel 165, because you can't bond to it. But still, 12 is enough to, you know, in a lot of cases, you can use 40 megahertz channels. 80 megahertz is where it starts to get, you probably don't want to do it. In fact, we recommend that you don't use 80 megahertz channels.
You drop down to six channels, which is, you know, pretty quickly run into overlapping channels. But there are scenarios where it can work, but, you know, I wouldn't use it by default. 160 megahertz, there's only two channels. We're back into 2.4, right? 2.4, there's three non-overlapping channels. 160 megahertz, you have two. It just doesn't make sense to use these as well as there aren't very many clients that support 160 megahertz.
So, just don't do 160 meg. How about if you want to migrate? You know, you have an 11n network or you have an 11ac network. So, what, you know, what do you need to do? So, if you're, you know, if you're anyway placing new APs on the wall or on the ceiling, it's a great time to reevaluate your Wi-Fi design.
And there's two schools of thought. Do you have a coverage-based network or do you have a capacity-based network? If you have a coverage-based network, most likely you design this, you know, you have an 11n network. When you designed it, Wi-Fi was kind of like a nice to have.
You just had a few devices on. So, in that case, we recommend that you redesign for capacity, right? You know, do a proper AP design, hire someone, do it yourself to actually design the network. So, that's going to work, you know, work for you over the next five years as the network grows.
If you already have a capacity-based design, evaluate do you have users complaining or do you not have users complaining? If you have users complaining, very likely you should redesign for your capacity needs. If you don't have users complaining, then think about am I expecting my network usage to increase? Am I expecting more users to come on? Am I expecting Wi-Fi to be more prevalent than, you know, kind of the default access medium? If yes, then you should redesign for your capacity needs. If you don't have users complaining and you're not expecting demands to increase, then, you know, you're probably okay to do a one-for-one swap.
But we do recommend that you, you know, you kind of go through this flow and this doesn't apply, you know, it applies to most cases. You know, kind of go through this flow and figure out do you need to redesign? Should, you know, because anyway you're placing new APs in the ceiling, you're spending that labor money, it makes sense to reevaluate now instead of six months later, oh crap, this isn't working the way I want it and having to add, you know, additional APs anyway. So, when we say, you know, coverage and capacity, coverage is essentially deploying the bare minimum number of APs that it takes to cover an area.
A capacity design is you deploy more APs than you need for the coverage, but you deploy the necessary number of APs for capacity. And we base capacity on the number of clients, the types of devices, right? Are they 11n versus 11ac versus 11ax? Are they one spatial stream, two spatial stream, you know, whatever it may be? And what's the, you know, expected usage or applications? If you have a VoIP network, voice over Wi-Fi, you're going to design a little bit differently than you would if, you know, just general email usage or expecting users to be able to stream video over Wi-Fi, right? We need to understand what's in use so that you can design the network properly. And also, 11ac devices are still the majority today.
You know, we have two 11ax client types out there. 11ac is going to be the majority, you know, for the foreseeable future. So you need to design your networks for 11ac devices. That's why, you know, at the beginning when I said designing for 11ax, design just like you would for 11ac. The rules don't change. Fundamentally, you need to evaluate what your needs are and design for them. So what are some of the pitfalls of 11ax? And so, you know, full disclosure, I work for a vendor. This is a vendor session, but here's what you should know about this. This affects all vendors, right? AP stability.
New APs, they require new firmware, new code, which have their inherent instability. Chipset immaturity. So, you know, regardless of your vendor, the chipset that your vendor has chosen for the AP, whether it's the two main ones are Broadcom or Qualcomm, have inherent limitations today.
So many vendors have been shipping 11ax. Some vendors have been shipping for a year, but many until recently have either, you know, the last two or, you know, one month or two months have just now turned on their 11ax features or they have not yet even enabled them at all. So you could buy an AP and not yet have 11ax features.
That's because the chipset manufacturers have only recently begun providing firmware with support for some of the mandatory 11ax features such as OFDMA. Or some of the other ones like, you know, BSS coloring is kind of sketchy support this time on both client and the AP side. And I should point out, you know, I talked about with the 8x8 APs, but some of the APs will never fully support 11ax due to hardware limitations.
So just know that, you know, from a chipset perspective, there's still improvements that are being made to the software. We do have client driver issues, unfortunately, and the big offender is Intel. They have, if you have an Intel client with drivers older than, say, a couple months, you know, somewhere around the, you know, March-ish to May-ish timeframes, depending on the chip, those Intel clients literally will not see an 11ax network.
The workaround is to disable Wi-Fi 6 features, HD advertisement from the AP. Or they do have upgraded drivers, but if you have a large Intel population, it can take a while to upgrade those drivers. So just one thing to be aware of.
We haven't seen, you know, we haven't seen it be a huge issue so far in the deployments we've done, but it's certainly something to keep in mind. And then lastly is lack of clients. We're only aware of two shipping clients out there.
You know, the client ecosystem will grow, but it's very small right now. And, you know, like I said, we're at the very beginning of 11ax. And those two clients are Samsung S10 and the derivatives, right? S10, S10e, S10+. These are, you know, here's one. This is an S10e. It supports, you know, it's a two by two client, 11ax. It supports the, you know, HE data rates. It does support OFDMA. I've tested and done using OFDMA.
The second client type is a Intel AX200 chip. And I think that there are some laptops that you can buy it in. If not, you can buy it, you know, a la carte off of Amazon or wherever you buy your chips from, and you can insert it into most laptops. Dell is pretty good with them. Asus is pretty good. Lenovo's usually you have problems.
You know, they locked down the bio, so you can't actually put it in the new chip. But if you want to test out 11ax, this can actually be a pretty easy way to test 11ax if your laptop has an M.2 interface. Again, the Intel AX200, we've seen the AX features work. We've seen HE data rates. We've seen OFDMA. They both seem to work pretty well.
How do you check, you know, if you're, you want to validate that an AP is using Wi-Fi 6 or 11ax? There's a couple of ways to do it. If you're, if you have a Wi-Fi analyzer, such as, you know, Wi-Fi scanner, Wi-Fi Explorer, Wi-Fi Explorer Pro, whatever, you know, channelizer on Windows, or yeah, Wi-Fi scanner, most of them will pick up the 802.11ax or Wi-Fi 6 from the beacon and actually display it in the, you know, in the interface itself. Now, if you have a newer Mac, there's a caveat that you have to do a passive scan, not an active scan. But again, you can pretty easily see if the APs are doing Wi-Fi 6, which is cool. You can take a PCAP. HE11ax will be advertised in the beacon.
You'll see HE capabilities within the beacon itself. And some devices will even tell you that they're using Wi-Fi 6. These, these Samsung devices, the S10s, when you connect to the Wi-Fi 6 network, it puts a little 6 next to the Wi-Fi symbol, which is pretty cool. And just kind of to wrap up, this is my last slide.
When should you migrate to 11ax? So after all this, you know, this is, this is a pretty standard technology adoption curve. We are at the very beginning, right? We are in the early adopters phase. We do have plenty of customers that have adopted 11ax, but these customers are typically risk tolerant, right? There are going to be speed bumps along the way, and this is true for every vendor.
Every vendor is going to have their problems over the next three months, over the next six months. It's going to take a little while to work them all out. And so if you are risk tolerant, you have a, you know, a refreshed need, then by all means, today is a great time to go 11ax.
If you are a little more risk averse, you want stability, you know, you can't handle downtime, then I would recommend waiting, right? How long you wait depends on how risk averse you are. I would imagine that within the next three or six months, you know, the majority of bugs will get worked out, and after that, you know, pretty safe to upgrade. But again, it all depends on your aversion to risk.
And so that's, that's what I have to say about just upgrading. It's, you know, AP, you know, APs are available, clients aren't really available. It's just, it just comes down to, you know, timing and your risk tolerance. So that's, that's it. I would like to thank everyone for joining. And I guess I can take questions if there are any questions have come up.
Mary Kate, I'm happy to take questions. Doesn't AX actually have worse performance than AC when there is only one client? No, I would expect AX to have better performance than AC because the maximum data rates, right? I mentioned that there's a different sub-carrier spacing. And so, you know, if I have, you know, my Samsung device versus my iPhone device, both are to spatial stream, I would expect higher throughput on my Samsung 11 AX device because of that, because of that data rate, about 20% improvement in data rate.
How realistic is actually using 80 megahertz channels? I would recommend against using 80 megahertz. You know, really, there's very few cases that you should be using 80 megahertz. You know, if you have high throughput demands, but low actual capacity, right, you don't have a lot of APs in one area, then you might be able to get away with 80 megahertz.
Or if you're in your house, you might be able to get away with it. But typically in enterprise, 40 megahertz is the widest that we go. And almost, you know, a lot of times we actually go down to 20 megahertz.
Let's see, for 11 AX, what are the known common interference elements that you will encounter in the real world? So, for 11 AX, I think, you know, really client adoption is the number one thing, both in terms of interoperability. Like I said, the only interop issue that we're aware of is with Intel clients, you know, with older drivers. But there is potential that other clients could have issues.
We just haven't seen them yet. Okay, well, we have a lot of questions. So, how mature is the coloring discrimination algorithms? They're very immature at this point.
You know, across the board, there has not been a lot of optimizations that have happened for BSS coloring. From a MIST perspective, this is an avenue that we want to pursue aggressively. We think that there's a lot of benefit to BSS coloring and potentially a lot of optimizations that can be done.
But we're kind of at the very beginning, you know, our chipset to be able to support it, as well as client devices to be able to support BSS coloring. And BSS coloring is an optional feature for Wi-Fi 6 certification. Any comments on 6 gigahertz? No comments on 6 gigahertz.
It will require, you know, upgraded hardware, but nothing to talk about at this junction in time. Is there an automated way for the switch or AP to determine the maximum bandwidth channels, maximum channel bandwidth available in a particular environment? From a MIST perspective, we do not dynamically choose, you know, maximum channel bandwidth. But from a technology standpoint, from an 802.11 standpoint, there is a feature called dynamic channel bandwidth allocation.
Most vendors have turned it off. We've turned it off. It's a little too finicky. And some other vendors actually support a dynamic bandwidth support, which I have found to be finicky, as well. But, you know, there are options to make that work if you need it. Do you feel client manufacturers will enable 11AX and 2.4? Yes, I do.
From the two devices that I know of, the Samsung as well as the Intel both support 11AX and 2.4. It's not much effort for these guys, so I would expect, yes, most devices will support 11AX and 2.4. I am not expecting many 2.4 only AX devices, but I suppose it is a possibility. Do APs support SD-WAN or SDN-WAN? Our APs will work in SDN, you know, SD-WAN environments, but we don't support SD-WAN natively. You would need a router for that.
How much CPU horsepower increase is required on AX clients? So, from an AP's perspective, most APs have significantly increased their horsepower. I'll say our AP, I think most other APs have gone from a dual-core CPU architecture to a quad-core. You know, there is potential for a higher packets per second where you need that CPU processing, but really, the clients just need to be – as long as they can understand the OFDM transactions and, you know, keep track of that kind of state, then they probably don't need – you know, clients themselves probably do not need that much more horsepower, but it's likely good on the AP itself to be able to keep track of all the different advanced features that are happening within 11AX.
When do you think Wi-Fi 6 will be the stable norm? I kind of talked about this. I'm expecting most APs to – three to six months is pretty safe, I think, in terms of AP stability across vendors. You know, it could vary on vendors. Certainly, from a MIST perspective, we're hoping it happens sooner. You know, from the deployments out there that we've seen, the APs – the AP works. You know, don't get me wrong there.
It's just – you know, there are certain nuances that – you know, we haven't really seen major client-affecting issues, but there are some things, you know, from a vendor perspective that we fix, and so it's – you know, for our customers, it hasn't been a huge issue, but for some customers, they never want to upgrade or have to upgrade, and now it's just not that time. If you deploy 11AX, expect, you know, firmware upgrades, you know, from your vendor over the next couple months. Is 11AX backwards compatible with 11AC? I can't believe I didn't mention this.
Yes, 11AX is completely backwards compatible with 11AC, and so an 11AX client, 11AX AP can – you know, 11AX client can connect to an 11AC network, and 11AC, 11N, you know, 11G, whatever, can connect also to 11AX APs. You know, no backwards compatibility issues there, except for that one Intel issue that I mentioned with the driver versions. So, save battery life, right? So, okay, this is a feature called TWT or target wait time.
I mentioned it briefly, but really, it's an enhancement over universal power save in 11AC. It just – it helps – it tells clients more prescriptively when they should wake up so that they, you know, can go to sleep longer, more frequently, and have, you know, save their battery life by only waking up at certain points to receive data. What happens if FCC gives 6 gigahertz channels with current APs? So, current APs across the board will not support Wi-Fi – sorry, 6 gigahertz.
It will require a new hardware, both on the AP side and client side, to be able to support 6 gigahertz. Can AX be powered over one cable with newer PoE, AF, BZ, BP? Yes. AX APs, there's – most AX APs will operate in some function over AT. Some of them will even operate under AF. Some require BT power. So, like, the MIST-AP43 operates fully under AT power.
Other vendors, you know, they have limited functionality under AT power, so it just depends on the AP out there. Where do you stand between 5G and Wi-Fi 6? This is a good one. My perspective on 5G versus Wi-Fi 6 is that Wi-Fi 6, you know, 802.11 Wi-Fi 6 is going to continue its life, right? It's going to be very popular.
Fundamentally, at the end of the day, it comes down to control – data control, right? In Wi-Fi, it's completely unlicensed spectrum. Whoever deploys the Wi-Fi – so, if a company is deploying Wi-Fi, they have complete control over what happens over Wi-Fi, the data that flows over that Wi-Fi. With 5G, that control shifts to the ISP or whoever is the provider, whoever is 5G network that is.
So, fundamentally, I believe the issue is data control, and for that reason, Wi-Fi 6 and Wi-Fi in general will continue to grow, continuous popularity. Asking about the AP43, which is the MIST AP, I haven't talked that much about it, because this is meant to be a more neutral webinar, but the MIST AP43 is our 11AX AP. It's 4x4 dual 5 capable. It has a 2.5 gig uplink. It has our Bluetooth antenna array. It's a pretty snazzy AP.
I encourage you to check it out. There's my one pitch. I just deployed 802.11ac wave 2 for a client.
Can I have a mix of 11AC and 11AX APs? Technically, yes, you can. Would I recommend it? Probably not. It all depends on the transition barrier, right? So, if you have – typically, the recommendation is within a building, within a floor, you have one model of APs.
It makes your troubleshooting so much easier. It eliminates – not that there will be interop issues, but it eliminates any potential interop issues. So, I would say that if you want a mix of 11AC and 11AX APs, make sure that there's a clear barrier, right? So, if it's between floors, have one model on one floor, another model on a different floor, or if it's between buildings, you have 11AC in one building, 11AX in a different building.
Do you think sports stadiums will look forward to Wi-Fi 6? Yes, I do think Wi-Fi 6 will make a difference in stadiums because of OFDMA and BSS coloring. It's going to take a little while because we're going to need clients, right? We need 11AX clients. It's taken – we've had 11AC for the past probably five years, four years we've had 11AC, and it's taken this entire amount of time to get to 90 percent, 80 percent client adoption of 11AC devices.
So, we can expect a similar adoption rate timeline for 11AX. So, it's going to take a little while before we actually realize the full benefits of 11AX because we need clients. Does the MIST-AP have its own DC jack if POE infrastructure can't be implemented in time? Good question. Let's see. Yeah, so, here's our AP, and it does have a DC jack. This is the AP43.
Let's see. Will the MIST-AP come with replaceable radios for future compatibility such as 6 gigahertz? I've answered this one a couple of times. No, it's going to require a hardware upgrade.
So, you're going to need a new AP for when 6 gigahertz comes around. What speeds have you reached on your Samsung device while on AX? I've seen over 500 megabits per second on the S10. I haven't really focused so much on single client, but 500 megabits for a two-spatial stream client is pretty good.
I think I've seen even close to 600. Actually, I have seen 600 with the S10. So, yeah, it's pretty good throughput. The device actually performs pretty well.
Wi-Fi 6(802.11ax)の概要
Wi-Fi 6を採用することで、ネットワーク容量、効率性、デバイスのバッテリー寿命が大幅に向上します。Wi-Fi 6の特徴は次のとおりです。
-
OFDMA:直交周波数分割多元接続(OFDMA)は、効率を高めるWi-Fi 6の重要な機能です。ワイヤレスチャネルを多数の小さなサブチャネルに分割し、各サブチャネルが異なるエンドポイント向けのデータを伝送します。この手法により、複数のクライアントへのデータの同時送信、遅延の低減、および帯域幅の使用の改善が可能になります。
-
BSSカラーリング:BSS(ベーシックサービスセット)カラーリングは、高密度のWi-Fi環境で重複するBSSの処理を改善する方法です。各 BSS に異なる識別子 (色) を割り当てます。これにより、アクセスポイント(AP)とクライアントは、他の BSS からの送信を区別して無視することができ、ネットワークの全体的な効率が向上します。
-
1024 QAM—直角位相振幅変調(QAM)が、以前のWi-Fi標準である256 QAMから、Wi-Fi 6では1024 QAMに拡張されました。1024 QAMでは、各信号がより多くのデータを伝送できるため、全体的なスループットが向上します。ただし、有効にするには、より高い信号対雑音比(SNR)が必要であり、範囲がわずかに狭くなる可能性があります。
-
アップリンクMU-MIMO—Wi-Fi 6では、アップリンクマルチユーザー複数入力複数出力(MU-MIMO)が導入されています。以前のWi-Fi規格では、APから複数のクライアントに同時にデータ送信できましたが、Wi-Fi 6では、複数のクライアントからAPへの同時送信もサポートすることで、これが改善されています。
-
ターゲットウェイクタイム:この機能は、デバイスがウェイクアップしてデータを受信するための所定の時間をスケジュールすることにより、デバイスのバッテリー寿命を延ばし、長時間アイドル状態を維持できるようにします(バッテリーを節約するため)。
Wi-Fi 6に移行すると、ネットワークの容量、効率、接続されたデバイスのバッテリー寿命が大幅に向上します。