Basics 3 Channel

– We are going to review this slide. Now, we talked about data rates early on, and we talked about– we talked about some of the legacy data rates that you’ll see. So we talked about 802.11b data rates. Remember that 1 and 2 megabits per second. These were the two original data rates that came out in the original 802.11 standard in 1997.

And then in 1999, we added 5.5 and 11 megabits per second. And then in 99 as well with 802.11a, we added the OFDM data rates, 6, 9, 12, 18, 24, 36, 48, 54 megabits per second. And then those were brought back down to the 2.4 gigahertz band with 802.11g. I’m sure you remember all that.

Then we talked about what 802.11n and then 802.11ac brought us which is what we call the MCS index. Now the MCS index is this chart that we’re looking at, and these are all of the possible data rates that we can get with 802.11n, those are all in red and 802.11ac, those are all in this slide blue color.

And there’s a lot of different factors that can– basically, a whole bunch of different combinations that can give us different data rates. We get different data rates depending on which factors we have in place when we configure our wireless network. For example, one of those things that can drive what kind of data rates we can hit are spatial streams.

We’re going to talk about this today. We’re going to talk about MIMO, what exactly MIMO is, why it matters, how it works, and some things that you need to be aware of, especially when it comes to access points. Things that you need to watch out for when you’re speaking out antennas and things like that.

So for example, if we take a single stream device like my very, very cheap Nokia 6.1, by the way, this is going to get replaced with an iPhone SE today. I’m going to order one after we are done here. But this is a single spatial stream device.

And so if we couple that with a 20 megahertz wide channel, the maximum data rate that we can hit with this device is 72 megabits per second. That’s as fast as it goes. Now on the 24th when I get my shiny new iPhone SE, I’ve been an Android user for nine years now, so switching to iPhone, this is a big deal.

That’s why I’m bringing it up here, is a two spatial stream device. And because of that, we can hit faster data rates. For example, we could hit potentially 144 megabits per second. Now my MacBook right now, it’s on a 40 megahertz wide network right now, and it’s a three spatial stream device.

And so in theory with 802.11n data rates, we should potentially be able to hit 450 megabits per second. That’s total theory. We typically we typically don’t get quite there, but we can hit fast data rates when we have more spatial streams available.

So I just want to really quickly review what are all the different things that can influence a data rate. What are the things that can influence the data rate? Well spatial streams. How many spatial streams do we have available? What kind of modulation are we using? That influences data rate quite a bit.

What channel width are we using? Is it a 20 megahertz channel? Is it a 40 megahertz channel? That influences data rate as well. We can also influence, it’s heavily influenced by what Fi type we are using. Fi type being 802.11b, 802.11g. In this case, it’s going to be either 802.11n or 802.11ac.

They are two different Fi types, and so they have different capabilities that give us different data rates. Also whether we’re using a long guard interval or a short guard interval. That one is probably the one thing that I’m going to leave out today. But I think the short explanation of that is that a guard interval is a little space.

It’s a little piece of empty air basically between transmissions. And with 802.11n, we introduced a shorter version of that, so we don’t take as long of breaks between transmissions. And that allows us to fill up the air with more data and thus hit faster data rates.

So we’re going to talk first through this modulation part. I want to talk about this a little bit and talk about exactly what these mean and why they matter for a wireless network design. Now just like last time, I want to remind you that you can get these charts. There’s two places you can get them.

You can go to mcs-index.com. That’s Mike, Charlie, Sierra, index.com. This particular chart came from our good friend Keith Parsons over at wlanpros. And so I would recommend having a copy of these on hand, bookmark this, something like that. Because I refer back to these all the time, and I think that you should too.
Basics 3.2
– Like I said last time, the ultimate goal that we have as wireless network engineers is we’re trying to go this way as much as possible. We’re trying to help clients get to the fastest data rate possible, and there’s a few reasons for that. One of those, of course, is because we want to deliver high-speed networks. We want to be able to move a lot of data as quickly as possible. We want the end user experience to be good. And to do that, we want to deliver fast data rates. We want to get this thing as close to ethernet as we possibly can.

We’ve never going to get there, but we want to try to get it close to ethernet, as close as we possibly can. We also want to conserve air time. I’m sure you remember from part one that on any given channel at any given time, only one device can talk. One thing gets to talk on a channel at a time. And so we want to try to give devices a way to get on the air, get off the air very, very fast. And the faster data rates we hit, the more air time we concern, everybody has a good time. We want to do that.

OK, cool. So let’s go ahead and move on here. Apologies if things seem a little bit clunky today. This is not my typical tool set, but we’ll work with it. Now, before we get into talking about modulation schemes, exactly how modulation works, I want to very quickly take a look at how ethernet and Wi-Fi are different. This is a little bit of review. We talked about this in part one, but I think it’s really important to understand how these two are different and to think about how do we make these faster over time.

Now, remember that Ethernet uses carrier sense multiple access collision detection. That’s really, really important. On the ethernet side of things, we can actually detect collisions. Over on the wireless side, we can’t detect anything. A wireless transceiver is either listening or transmitting. It can’t do both at the same time. And so we have to use carrier sense multiple access collision avoidance. Basically, the whole protocol is designed around how to avoid collisions. Now, an ethernet NIC only receives the frames that are destined for it.

If you have it connected to a switch, we’re only going to get frames that are meant to go to that NIC. Unless you’ve got an old hub or something like that, it’s only going to get frames that are destined for that device. Whereas a wireless NIC has to demodulate everything that it hears. If it hears something, it has to demodulate that frame and look at the MAC address and go, huh, was this meant for me or not? Only then can it throw the frame out. So it has to demodulate everything that it hears.

On the ethernet side of things, we see a consistent data rate. Things are usually very consistent. You plug in a cable and you usually get 1,000 megabits per second. It’s typically like that. Not always, there can be some problems there. And we’ve done some really cool stuff on the wired side of the network that I’m really excited to talk to you about today. Definitely want to spend some time talking about wired assurance and what we’ve done over there with Juniper EX switches. There’s some fun things to talk about there. But for the most part, we see very consistent data rates.

Over on the wireless side, the data rate varies wildly based on signal strength, signal to noise ratio, how much traffic there is on the channel. There’s a whole bunch of factors that affect how fast we can go on the wireless side of things. Over on the ethernet side, throughput is the line rate. If we have a 1,000 megabits per second available, depending on what we have throughout that transmission process, we’re going to be able to hit that at phi level. We’re going to be able to hit 1,000 megabits per second.

Whereas over on the wireless side, the actual throughput is roughly 1/2 the data rate. So that means that if my little Nokia 6.1 is pulling 72 megabits per second, well, we’re going to be lucky to see 40 or 50. That’s the reality is that the throughput is roughly 1/2 of the data rate. Be sure that in your mind, you’re separating throughput and data rate. They’re two different things. They’re linked, but they’re not the same thing at all.

So also keep in mind that throughput decreases as the number of clients on the channel increases. I think that the reality of modern access points is that we don’t really run out of access point today. Maybe one of the hardware developers or former engineers or something like that, maybe they would tell me that I’m wrong on this one, but I think that modern enterprise grade access points have tons of CPU, tons of compute power, and they can handle a lot of clients from that perspective.

But what do we run out of? We run out of channel. We always run out of access to the medium before we run out of resources on the access point itself. Many access point vendors out there, you look at the spec sheets and they say that they can support 512 clients per radio. That’s a fairly common number that I see quite a bit. But the reality is that depending on your usage, you’re probably not going to clear more than a couple of hundred client devices per radio. And it depends wildly on what you’re using those devices for.

If you’re doing VoIP calls, for example, it depends, but what are we talking about? 20 devices? 10 devices? It really depends. Whereas if we’re talking about IoT devices, it could be a couple of hundred very, very easily. So it wildly depends on what we’re using the network for, what we’re using the channel for. We always run out of channel before we run out of access point hardware. So to kind of wrap this discussion up, what I want to do is think just for a second about how we make things faster over time.

If you look at ethernet, how did we make ethernet faster over time? Did we put better copper in the cables? Is this better copper than it used to be? No, that’s not what we did. We decreased interference. The way that we made ethernet faster is by decreasing interference. We did some things like we added shielding to the cable to keep outside interference out. We engineered a very specific twist rate for our pairs of copper so that there’s no crosstalk. We have reduced interference, and that’s how we made ethernet faster.

So how do we make Wi-Fi faster? Well, the way that we make Wi-Fi faster is to reduce interference, thus, we can push faster data rates. So let’s talk about that. Let’s talk about that a little bit now. So let’s see, next slide time. So before we can get into talking about how we need to reduce interference, before we can talk about what we do to make Wi-Fi faster, we need to spend just a couple of minutes on modulation. We talk about different types of modulation that are out there.

So I just want to walk through those really quick. We’re going to start with baseband, and then we’ll talk about amplitude shift-keying, then we’ll talk about frequency shift-keying, and then phase shift-keying. So basically, four or three different types of modulation starting with baseband. So baseband modulation is where we take 1’s and 0’s. And so let’s do that really quick. Let’s just make up some numbers here, 00101101.

There we go. And, oh, totally by accident, when I made this slide, I actually made a complete byte out of that. I didn’t mean to do that, but that works out really nicely. So with baseband, baseband as a kind of form of modulation, we’re just looking at whether we have an on or off, is this a 1 or a 0? So for example, a 0 would be low, another 0 would be low, a 1 would be high, a 0 would be low, and then we’d have high, high and then low, and then to finish it out, we’d have one more high. That is baseband.

That is a baseband modulation because it’s either high or low, and that’s it. If it’s high, it represents a 1. If it’s low, it represents a 0. But how do we put this across analog? So what we have to do is we have to put this data across an RF wave somehow. You remember this. We talked about RF waves. We talked about 2.4. We talked about 5 gigahertz. So here we have an RF wave.

And you can see it going through all of its cycles here as it starts at 0 degrees, goes to 90 degrees, goes to 180 degrees, 270, and then 360 degrees as it completes its entire cycle. We have to take baseband, and we have to put it on this analog waveform somehow. So there’s a few ways that we can do that. One way that we can do it is with amplitude shift-keying. So if somebody wants to police the mute, that would be awesome. Thank you very much.

With amplitude shift-keying, we are going to change the height of the wave. And so if we want it to be a high-amplitude wave, the wave is going to be really tall like this. And if we want it to be a low-amplitude wave, then it’ll be a really short wave like this. That’s changing amplitude. In fact, you have probably been using an amplitude modulation system for years and years and years that you didn’t even know about.

Amplitude modulation is how we put voice on a radio wave. So when you’re listening to AM radio in your car, the way the AM radio works is it changes the height of that sine wave, and the height of that sine wave carries a frequency range that you can hear. And so the radio takes that and it pulls that secondary waveform off of that amplitude modulation wave, and there you go, now you have amplitude modulation that’s giving you something that you can hear.

So amplitude shift-keying is a little bit different because instead of using it to carry an analog signal like audio, it’s going to carry a digital signal. How do we do that? Well, if it’s a 0, we’re going to see a very low amplitude. So we got another 0 here. So we’re going to have more low amplitude. And then to carry high amplitude or then to carry a 1, we will switch to a high amplitude wave. So you see what just happened there? That wave got a lot taller. We just put a little bit more power into the wave, and it represents a 1 now.

So we go back to a 0, and then we get back to a 1, and we have something tall. And you can see how that works. Pretty clear, right? Pretty easy. So that is amplitude shift-keying. So next is– let’s see. I’m going to go hit the attendee list just a little bit here and take care of a couple people that are new. There we go, got you now. You’re all good. So now let’s switch over to frequency shift-keying. Now, the way that we use frequency shift-keying is we change the frequency of the wave.

I wonder if I can– oh, yes, I can just– oh, I erased it all. That’s not what I meant to do. Oh well, I think you all have a picture in your mind of what it looked like. So let’s just redo our– we’ll do 01001101. And so I think Cydia is going to have to revise this statement that I have mastered annotations in Zoom and RingCentral. I have not mastered annotations in PowerPoint. It’s all about using the right tools.

So with frequency shift-keying, we’re going to change the frequency of the wave. Now, you remember we talked about 2.4 versus 5 gigahertz. You remember that we talked about how a 2.4 gigahertz wave, it’s a longer wavelength, whereas 5 gigahertz is a shorter wavelength. Those wavelengths oscillate much faster. They’re much closer together. So we can look at the length of a wave, how long it takes it for it to complete its cycle.

So here’s the beginning of a cycle. Here’s the end of the cycle. That is our wavelength. For a 5 gigahertz one that’s a higher frequency, so we would say this is the beginning of our cycle, this is the end of our cycle. So it is a shorter wavelength. We can carry 1’s and 0’s. In fact, we can carry lots of different types of information by also changing the frequency of the wave as we transmit it.

Now, one technology that you have used before as well, I’m absolutely sure that you’ve used– OK, so if we go to eraser, there we go, I’m learning how to do this, learning how to use the annotation tools on the fly here –you have used a type of frequency modulation before when you’ve listened to FM radio. The way that FM radio works is it modulates the audio by changing the frequency of the wave on the fly. And so, for example, we can see that here we’ve got a longer wavelength.

And then we have this period here that has a shorter wavelength. And then we go back to a longer wavelength again, and then back to a shorter wavelength again. And so you can really easily see how we could carry a secondary wave on this. We could easily carry some kind of audio wave by modulating the frequency. That’s an analog way of doing it, but now we’re going to look at the digital way of doing frequency shift-keying.

So let’s say for a 0, let’s do a long wave. And so a long wave represents a 0. And then if we have short waves, that represents a 1. Then we go back to a long wave again and a long wave again, and then some short waves to represent a 1. And you see where this is going. You see how we’re changing the frequency to say whether it is a 1 or a 0. So just to kind of fill this back in, let’s do short waves.

It’s going to be a little bit sloppier this time around because I don’t want to spend too much time just redrawing things. And then we’ll go back to a taller wave. And I think that’s enough bits for now for you to understand what we’re talking about. OK, so we’ve talked about baseband. We’ve talked about amplitude shift-keying. We’ve talked about frequency shift-keying. That leaves phase shift-keying, the last most difficult to understand, but probably the most important one when we’re having a discussion about Wi-Fi and how Wi-Fi works.

So let me get rid of this really quick so we can have a little bit more space. Phase is a comparison between two waves. And so let’s say that we take one radio wave. And so here is one sine wave here. It starts and it ends. Phase is a way to measure how offset these two waves are. So for example, if we start another wave right here and it just kind of follows along and finishes up a little bit later, when we compare these two waves together, the second wave is said to be 90 degrees out of phase.

That’s because with a wave as a wave goes through all, as it goes through its cycle, there are basically different degrees that it goes through. So think of it kind of like a circle. You take a circle, you have 0 degrees, you have 90 degrees, you have 270 degrees, and then you have 360 degrees. And as the circle goes around, you can assign degrees to that. Well, a wave is the same thing. Think of a wave as a circle.

When you look at a sine wave, it’s just the same thing repeating over and over and over again. It’s a wave going through a loop. You kind of have to tie those two things together in your mind to understand how this works. So when we look at a wave, we’ve got 0 degrees, then it reaches 90 degrees, then it reaches 180, degrees then 270 degrees, and then finally 360 degrees, and the cycle repeats. When we look at two waves, when we compare them, we can compare them in degrees.

How far shifted are they from each other in degrees? And in this case, this wave has a difference in phase of 90 degrees. That is a 90 degree difference between these two waves. Now, where do we see shifted phases in the real world? Where are some places that we see it? Well, think about the room that you’re sitting in right now. Are there any echoes in your room? Maybe you have a big rug or something like that. I’ve got some of these really cheap sound absorbing panels from Amazon.

The only thing that those do, they don’t actually keep the outside world quiet. If my kids are playing over in the next room, and I’m sure your kids are doing the same thing, they’re all playing over there right now, making a bunch of noise and stuff. You just get to be muted. I don’t. But these panels don’t keep outside sound out at all. What they do is they reduce echo. So if I clap, you can probably hear just a teeny tiny bit of echo.

There was a lot more echo before I put these up. What echo is is echo is when my voice leaves my face and reflects off of all the different surfaces in my office and then comes back to my microphone slightly out of phase with each other. And so what you’re getting is you’re getting multiple copies of my voice that are all arriving back at this source at slightly different times, and that’s what causes echo. And so we’re going to talk about that a little bit more in just a couple of minutes.

But that is phase shifting happening in the real world, in the audio world. That’s something that you can observe, something you can hear everywhere that you go. Maybe you’ll start listening for it now whenever you’re in an interesting room with neat geometry or lots of hard surfaces or something like that. So with phase shift-keying, what we’re able to do to modulate a 0 or a 1, let’s just draw a normal sine wave here. So there’s our normal sine wave. It starts at 0, ends at 360.

With phase shift-keying, we can change when that sine wave begins and ends. And when we shift that sine wave back and forth and shift it, have it start and stop at different intervals of its process, we can apply that as modulation. Now, I’m doing a lot more than just 1’s and 0’s right now because I just kind of want to give you the full effect of what this might look like. But you can see where it’s not a consistent sine wave going through its cycles, it’s changing phase.

We’re moving the phase forward and backward to convey different pieces of information. OK, cool. So let’s move on, if I can figure out how to move on. All right, cool. So now we can kind of start getting into how modulation works. And so what I’m going to do is I’m going to talk through the simplest type of modulation that there is, and this type is called BPSK or Binary Phase Shift-Keying.

So I just went back to Zoom and tried to use the annotation again because it’s habit, been doing it that way for years. And so this is going to be very challenging throughout the day to remember, nope, got to use the PowerPoint annotations. Come back. There we go, Pen. And now we have to go back a slide. BPSK stands for Binary Phase Shift-Keying. With binary phase shift-keying, we are going to represent a 0 or a 1.

This is an extremely simple form of modulation that gives you very slow data rates. You get 6.5 megabits per second out of binary phase shift-keying. So how do we do binary phase shift-keying? Well, let’s start by drawing a sine wave. So I’m going to draw a sine wave here. And it starts at 0, and it comes up at 180 degrees, peaks at 270, and then comes back down and stops at 360 degrees.

So when we transmit a sine wave like this, just a normal sine wave, that is going to be a 1. If we transmit a minus sign, so the opposite of this sine wave, that would be a 0. So it depends on which wave we transmit, whether it’s a 1 or a 0. And you can see how that’s using phase shift-keying. It’s moving that phase forward or backward to represent a 1 or a 0. Now, there’s another way we can represent that, and that’s with a polar coordinate chart.

With a polar coordinate chart, we’re going to take this same information, and we’re just going to map it in a different way. So we’re going to say that 0 degrees is up here, 90 degrees is over here, 180 degrees is down here, 270 degrees is over here. And there we go. Now we have our polar coordinate chart. I’m going to go ahead and erase one of these really quick, race pin. And we erased it all.

This is driving me crazy. I was already rattled because of the reboot issues and stuff, but this is not great. It’s not going super well. So let’s draw our 1. So this is going to represent a 1. So here’s our normal sine wave. If we draw that on a polar coordinate chart and we look at, where does this sine wave peak? It is peaking right here at 90 degrees. That’s where it peaks.

And if we want to assign that to a polar coordinate chart, we’re going to start right up here at the top at 0 degrees and we’re going to go until we peak, and that peak is right here at 90 degrees. And so when we look, where does our sine wave peak? It peaks right here at 90 degrees. Now, if we look at our minus sign, where does that peak? So let’s draw that really quick. So we’ll go to 180. It peaks at where? 270 degrees.

So if we look at that on the polar coordinate chart, that comes around here and it peaks right here at 270. So there we go. So when we look at this on a polar coordinate chart, you see how we can basically see where does the target land. Depending on where does this sine wave peak, where does it land in our polar coordinate chart? And depending on which side of the chart this lands on, that tells us is it a 1 or is it a 0.

If it lands on this side of the chart, that is a 1. If it lands on this side of the chart that is a 0. This box is called an EVM box or Error Vector Magnitude. And what’s great about this box is that it’s big. It’s a really, really large box, which means if this sine wave gets messed up a little bit, we can still understand whether this was a 1 or a 0. So let’s take a look at that scenario really quick.

So let’s say that when we intend to transmit a 1, but the sine wave gets a little bit garbled up and it’s kind of hard to tell, OK, where is our peak at, and it really ends up looking more like it’s at 45 degrees. That’s where it looks like it peaked at. So the radio on the other end goes, OK, well, I think it peaked at 45 degrees. And so that would be right about here. Which EVM box did that land in? Well, it very cleanly landed in a 1.

And so even though our sine wave was not clean, even though the receiver on the other end could not tell exactly, even our transmitter couldn’t get exactly what it wanted to across the board here, we were able to land that in our EVM box of 1, and, thus, we’re able to demodulate that as a 1. All we have to do is make the target either land on the right side or make the target land on the left side, and we can tell whether it was supposed to be a 1 or a 0.

OK, so how do we make things go faster? Well, the way that we make things go faster is we make our modulation more complex. So let’s do that. So what I’m going to do is let’s go take a look at the next type of modulation, which is QPSK or Quadrature Phase Shift-Keying. So with quadrature phase shift-keying, instead of just transmitting a 1 or a 0, we are going to transmit two bits at a time.

So we can send instead of a 1 or a 0, now we double the amount of data that we can send in the same amount of time. We can now send two bits at a time a 00 or a 01 or a 10 or a 11. And so the way that we do this is we make that sine wave peak at different points throughout our polar chart to say whether it’s a 00 or a 01 or a 10 or a 11. So let’s take a look at what that would look like.

So let’s say that it peaks at 45 degrees. If it peaks at 45 degrees, then that means that our target lands right here on our polar coordinate chart. So then let’s break our polar coordinate chart down to say what all these different parts of the chart represent. So this upper right-hand corner, I’m just going to put some numbers on here. I don’t know if this is exactly where they land for 802.11, but you get the idea.

This is just a teach you the concept of how it works, not where the actual bits and bytes are. So let’s say that this one is 00, this one is 01, this one is 10, and this one is 11. So if we peak at 45 degrees and we land inside 00, notice that our EVM box, our error vector magnitude box, just got 1/2 the size because now we can’t just land on either side of this polar coordinate chart. Now we have to land it in a corner of the polar coordinate chart.

If we land here that’s a 00. If that sine wave peaks closer to 180 degrees like maybe right here, that’s a 01. If it peaks very close to 360 degrees, that would be up here, and that’s going to be a 11. And so you can see how we are increasing the complexity of our modulation to get more bits through the air at the same time. So how do we make it faster? Well, we go for more complex modulation. So let’s do that.

So I’m going to erase it, see if we can move forward here. Hey, PowerPoint, want to move forward? There we go. And now we come to 16-QAM. And there’s a new element of complexity here. Hey, Power, erase. OK, so now we are going to up the game a little bit by moving to 16-QAM. So 16-QAM adds another element of complexity to our modulation. Now, so far, all we’ve been doing is shifting phase.

We’ve been moving that wave forward or backward to make it peak in different places, but what we’re going to do now is we’re going to add another layer of complexity to this by adding amplitude. Now, you remember earlier, we talked about amplitude modulation, how a low wave can indicate one thing and then a higher amplitude wave can convey something else. We’re going to mix phase modulation and amplitude modulation together now or amplitude shift-keying rather to create something new, 16-QAM.

So to do that, let’s erase our pen and start over here. So let’s say that we want to transmit a 000. Well, what we’ve done now is our EVM boxes have shrunk once again to a 1/4 of their size before. So you probably remember that with BPSK, our EVM box was 1/2 of the coordinate chart, and then with QPSK, our EVM box was now a 1/4 of the chart. So that was where our EVM was for QPSK.

Now for 16-QAM, now our EVM boxes are getting much, much smaller, but we have a new dimension to work with here. We don’t just need to worry about how far around the chart we are. We also need to worry about how far away from center we are to land in all of these different boxes to convey the correct bits that we want to convey. So how do we do that? Well, we use phase shift-keying to say, where does the wave peak? And then we use amplitude to say, how far from center do we want it to peak?

So for example, I’m going to clear these again. And this is a very mental exercise to clear these every single time. If our wave looks like this, then maybe our peak is right here. If our wave looks like this, then maybe our peak is out here or right here or right here. You can see how now we’re using both phase to change where around in the process we are and we’re using amplitude to say how far from center we are.

But the problem with this, the problem is that now our EVM boxes are getting pretty dang small. So I want you to think about this from maybe like– I don’t know. If you’re into firearms or something like that, think about it from a target practice standpoint, if maybe you had a firearm, maybe you’re a bow and arrows or something like that. As a kid, I had a bow and arrow. And if I stood a certain distance away from a hay bale, I could hit a reasonably small target, and I felt pretty confident about that.

But as I started to move away from that hay bale, it would be more and more difficult to hit that target. What if you made the target smaller? That’s what we’re doing with EVM boxes. As we go up in complexity in our modulation scheme, our EVM boxes get smaller and smaller and smaller as we go. But how far can we push this? Let’s see how far we can push it. So now let’s move ahead to 64-QAM.

Now, I’m not going to go through and worry about drawing in all the 0’s and all the 1’s, but now remember we were looking at 1 bit with BPSK. With QPSK, we were looking at 4 bits. And now all the way up at 64-QAM, now we’re looking at 6 bits. Sorry, with QPSK it was 2 bits, and 16-QAM, it was 4 bits. Now we’re looking at 6 bits worth of data that we can transmit at any given time.

These bits, if you look at these bits structures, these are called a symbol. I was going to try to write this out. I was going to try to write this out, but no-go, but these are each called a symbol. These are each called a symbol, a 6-bit structure, a 4-bit structure. We get different symbol sizes based on the modulation scheme that we’re using. OK, so 64-QAM, what happens with 64-QAM?

Well, let’s review really quick. Remember with BPSK, we had this massive EVM box. With QPSK, we had a slightly smaller EVM box. With 16-QAM, we had an even smaller EVM box. And then we move to 64-QAM, and our EVM boxes get teeny, teeny, tiny. And so what we’re asking our Wi-Fi transmitters and receivers to do is we’re asking them to understand, OK, which angle did our wave arrive in and we’re also asking them to understand what amplitude did we receive that way at to understand where we’re going to land in the chart to say which set of bits, which symbol are we transmitting here.

Holy cow, this EVM box is getting teeny, teeny, tiny, but wait. How far can we actually push this? Let’s find out. So with 802.11ac, we got 256-QAM. Now, I got really lazy here. I didn’t want to draw on all the lines for this because with 256-QAM, we’re going to double up the amount of lines. I’m not sure why PowerPoint moved back one slide just because I started to annotate. I’m not sure what’s going on there.

And it’s again. Oh, man, PowerPoint, you’re driving me crazy. So we have 64-QAM. We go to 256-QAM . Drew them all the wrong slide. Guys, this is just not going well. Things are not smooth today. I really appreciate your patience though. You’ve all been absolutely fantastic. I’m just going to fill in one of these. If we go to 256-QAM, look at how tiny our EVM boxes get. Now we’re looking at an EVM box like this, teeny, teeny, tiny EVM box.

But then now with 802.11ax, we get another modulation scheme. We get 1,024-QAM. With 1,024-QAM, now instead of transmitting 8 bits per symbol just like we were with 256-QAM, now we’re transmitting 10 bits per symbol. And what happens to our EVM boxes? Our EVM boxes are now this big.

So if you’ve ever done throughput testing back when you got your very first 802.11ac access point, you got your very first 802.11ac access point, maybe it was like a Netgear, like a home thing or something like that, and you managed to get your hands on an 802.11ac client, you probably noticed that you only got those awesome data rates when you were right on top of the access point.

If you were 10 feet away from the access point, yeah, no problem, you would hit these amazing data rates. But as soon as you started to move away from that access point at all, you saw your data rates go down and down and down and down. Why is that? It’s because of QAM. It’s because of the way modulation works. When we’re very close to an access point, when our signal strength is really good, when our signal to noise ratio is excellent, then our transceivers are able to decode, they’re able to decode these with great accuracy.

Every time a symbol lands within our EVM box, we can understand that. But as our signal to noise ratio starts to get worse, as our signal goes down, all of a sudden, the Wi-Fi receiver can’t quite place these in that EVM box like it wants to. Things start to get a little bit messy. So what do we do? We rate shift. We move back to a larger EVM box. So in this case, I’m going to move back to 256-QAM. And now we’re mostly able to get our symbols to land in the boxes that we want them to land in.

But as we start to lose signal strength, as SNR goes down, then we’re going to move back to a larger EVM box, and so on and so forth. We keep doing that until eventually we end up all the way back down at binary phase shifting-keying with a massive EVM box, but we could only send a 1 or a 0 at a time. So–

– Joel

– Yes?

– Joel, a quick question.

– Yeah, go ahead.

– For the 256-QAM, there are 8 bits per cycle and we sample the full sine wave at every 45 degrees.

– So yeah, good question. There are some of the real specifics of this. This kind of hits a point pretty fast where my ability to articulate what is going on drops off pretty rapidly. It drops off pretty rapidly. But what you need to know, I think the most important thing for you to know about this is that when a Wi-Fi receiver hears that sine wave, it is essentially trying to place a dot somewhere on this polar coordinate map.

And in any one of these boxes, this box might represent 0000000. This box might represent 0100000. This box might represent 000100. You see the pattern that I’m building here, right?

– Yes, of course.

– If that transmission lands in that box, we know it was a bunch of 0’s. If that transmission lands in this box, we know it was a bunch of 1’s. And what’s really important here, it’s very important to have good signal to noise ratio to make sure that we can accurately place these in the box. We’re running CRC checks in the background here to make sure that the data that we’re receiving was actually understandable. And when SNR gets bad enough, then our receiver will stop sending acknowledgments.

It’ll say, OK, that was a big mess, none of those 1’s and 0’s made any sense together, they didn’t pass their CRC checksum. And so I don’t understand that frame, and we do not send an acknowledgment. Does that answer your question at all? Or does that give you any additional insight?

– Yeah, that answers. One associated question is, so in order to be able to place that into this box, do we sample the amplitude at every 45 degrees?

– I’m not sure how often it’s sampled. I think you’d have to talk to a real wireless network engineer to know, somebody that actually works on the radio. I’m thinking somebody that works at Qualcomm or Broadcom and really deeply has an understanding of how the phi works. I don’t know the exact nature of how often it is sampled, but I think it has to be more than every 45 degrees because, to me, I see a lot more options than just every 45 degrees.

This might be every 15 degrees. This might be every 10 degrees. And so I think the sampling has to be very, very accurate.

– Understood. OK.

– Yeah.

– It’s a combination of amplitude and phase.

– Absolutely.

– The 256 comes out and it depends on different radio, how many amplitudes versus how many phases you’re actually using together in order to get the 256 different combinations.

– Yep, absolutely. So yes, very interesting topic. And I am explaining something that I think I have a decent grasp on, something that I would not say that I have a strong grasp of though. I think my understanding of this is good enough to explain the concept and to understand why we need good SNR, and why we design networks for a certain level of signal strength, and also why signal strengths fall off when you start to move away from the access point physically.

– Sure, Joel, you are doing great. Thank you.

– Thank you. I appreciate that. I like this topic a lot. I think it’s a lot of fun because it just kind of gets into some stuff that I don’t think most people ever even think about. I think most people, they never get to this point. And so I think it’s really fun to spend some time in it. Cool.

– Hey, Joel.

– Yeah, go ahead.

– There is one question. And maybe we can get to this. I think you’re going to talk about data rates, but Daniel asked, what mechanism is in place that changes the data rate?

– Yeah, great question. And normally, I keep an eye on the chat, but today, RingCentral, the background of the chat, is black and the text is black, so I can’t read anything unless I highlight it. So yeah, this is a nightmare. This is just absolutely awful. What else could go wrong? Watch the power is going to go out, and then my cellular failover won’t work. And yeah, that could be the only thing else that goes wrong today.

So the mechanism that changes the data rate is all based on retries. Daniel, I don’t know if you remember, I think it was in part 1, we talked about retries. And so there’s this constant process where we have a client device. So here’s my shiny iPhone SE that I’m going to order here later today. And then we have our missed AP over here. And the phone will transmit some data to the access point and the access point will check it to make sure it understood that data. If it did understand it, it’ll go, yep, check, and it will reply with an acknowledgment.

And so it’s this constant process of data acknowledgment, data acknowledgment, data acknowledgment. You can flip these arrows the other way, by the way. It doesn’t matter which direction this is going in. And this whole process is based on a retry rate, what percentage of the time do we have to retry, do we have to re-transmit that data? So what this handheld device is doing, what the iPhone SE is going to do here is it’s going to try a fast modulation scheme.

And if it doesn’t get a response, maybe it’ll try it again. And if it still doesn’t get a response, it’ll go, OK, well, let’s try 16-QAM. And so it’ll try 16-QAM, and it still doesn’t get a response. Oh, OK, so now we try BPSK. OK, BPSK worked. And so maybe we’ll hang out at BPSK for a while. And then maybe every now and then, we’ll test the waters and we’ll try a faster data rate. Let’s try 64-QAM. And the 64 QAM, hey, that worked. And so we get an acknowledgment back. We send another one. Hey, that worked. We get an acknowledgment back.

And so these chip sets are constantly looking at, when I’m transmitting to this device, how often am I having to re-transmit? If when my re-transmits hit a certain threshold, I am going to shift down to a lower data rate and, thus, down to a simpler, more basic modulation scheme. And so it’s that constant process of looking at acts or looking at acknowledgments and seeing how often we have to retry things. And by the way, that is 100% up to whenever the phone transmits.

It’s up to the phone. Whenever the AP transmits back to that phone, it’s up to the AP. They’re both maintaining these tables in their heads of, OK, when I talk to this device, what data rate do I need to use versus when I talk to this device based on how well things are going?
Basics 3.3
– Now we arrive at the SNR slide. Now we talked about the SNR slide earlier. We talked about this in either part 1 or part 2. I don’t remember. But we did talk about this one. But what I want you to remember about signal to noise ratio is this says how much signal is there above the background noise. So think about every room that you’ve ever been in has a certain level of background noise.

If you if you don’t have headphones on right now, you might hear my voice, the signal, right? And a pretty good sounding signal, If I don’t say so myself. But I’m kidding, of course. But you’re going to also hear background noise, right? The HVAC in your house, maybe there’s some cars out on the street. Maybe your kids are playing in the next room. That is all background noise. That is a noise floor.

There’s noise in the background. It’s true in an audio environment. It’s true in RF, as well. There’s always a certain level of background noise. And we need our signal strength to be a certain level above that background noise to be able to understand what other devices are saying. And how much SNR we need depends on the modulation scheme that we want to hit.

So, for example, if we want to hit 256 QAM. If we want to hit 256 QAM, we need to have lots of SNR, really, really good SNR. If we want to hit 64 QAM, yeah, we need quite a bit of SNR, right? And I’m just drawing some random lines here. If once we get down to a certain– when SNR starts to not get so great, we might have to go down to 16 QAM, we might end up all the way down at– we might end up all the way down at BPSK again, depending on how much signal strength we have above the background noise.

So this is why it’s really, really important when you’re designing a network, when you are sight surveying a network, when you’re looking at the service level expectations in the Mist dashboard, it’s really, really important to understand that you have to provide a certain level of signal strength, thus a certain signal to noise ratio, to hit these fast modulation schemes.

Without them, you’re going all the way down to BPSK, all the way down to 6.5 megabits per second. It’ll work great. It’s going to be super reliable. It’s just not going to be fast, it’s going to consume a lot of air time. Cool. Any other questions around that? Or are we pretty much good to go there?
Basics 3.4
– OK. Next thing I want to do is I want to talk a little bit about OFDM and how OFDM works and how all this stuff that we just talked about, we just took all of this information about modulation schemes and how that all works, and I want to apply it to OFDM to help you visualize how this all comes together.

Now, you remember I talked about OFDM. I’m sure you’ve memorized that OFDM stands for Orthogonal Frequency Division Multiplexing, right. You remember that. What you probably also remember is that OFDM makes a flat table top shape in the spectrum. If you look at it on a spectrum analyzer, it looks like that.

And that is a data rate of 6, 9, 12, 18, 24, 36, 48, and 54 megabits per second. That’s all of our 802.11g data rates. That’s also all of our 802.11a data rates, same data rates. This is also the modulation scheme that 802.11n, 802.11ac, and to a degree– I’m not going to get into it today– but to a degree also what 802.11ax use.

If you look at any of those types of traffic on a spectrum analyzer, any of those modulation schemes that we just talked about, they’re all going to look like this, that flat tabletop shape in the spectrum. Whereas old 802.11b stuff or 802.11 prime stuff makes a curve shape in the spectrum because that is a different modulation scheme called DSSS or Direct Sequence Spread Spectrum is what that stands for. Everything modern though, everything modern is OFDM. It’s all OFDM.

So what is OFDM actually composed of? If you were to zoom in to one of these OFDM signatures, if you were to zoom into one of these and look really, really closely with a very, very high performance spectrum analyzer with tons of resolution, what you would see is you would see PowerPoint change slides on you unexpectedly. That’s what you would see.

Now, so what you would see is you would see a whole bunch of little spikes called subcarriers. Each one of these little spikes is called a subcarrier. And if I remember correctly– oh, yeah, we’re good here– there’s 64 subcarriers in an OFDM signature for Wi-Fi on a 20 megahertz wide channel. Now, you bump up to a 40 megahertz wide channel, there’s more subcarriers. You bump up to an 80 megahertz wide channel, there’s more subcarriers. But for a 20 megahertz channel, we get 64 subcarriers.

Now, they aren’t all used for data all the time. Notice the red ones on the side. These are actually null subcarriers. We actually don’t transmit anything in those subcarrier spaces. Nothing gets transmitted right there.

There’s also one null subcarrier here in the center. We don’t transmit anything on that ever, ever. That one never gets lit up. In fact, when you look at an OFDM signature on a spectrum analyzer, you’ll almost always notice, even on a very coarse spectrum analyzer without a lot of resolution, you’ll always notice this little dip in the center. That little dip is that null subcarrier that’s not transmitting.

We also get pilot subcarriers. If we’re transmitting, these pilot subcarriers transmit constantly. They’re basically there to help align the receiving radio and help it understand exactly where in frequency space where this transmission is going to land. And that brings us to the last kind. That brings us to the data subcarriers.

These are where we actually carry the data. So think of data subcarriers as like strings on a guitar. I play guitar very, very badly. I know my bar chords, so that’s something. But I don’t have any calluses anymore. I don’t play enough. I don’t stay good at it. But I’m sure all of you understand that when you strum a guitar, you finger multiple notes on the guitar, and you strum multiple strings, and that produces a chord right.

These are all like strings on a guitar, a guitar with a ton, a ton of strings. For every single transmission, every time we transmit something, we put a symbol. Remember those? Remember we talked about symbols. We put a symbol on each one of these subcarriers at each one of these individual little frequencies.

So here’s the crazy thing about OFDM. Here’s the crazy thing about BPSK and 16 QAM and 256 QAM. And all of it is that it’s not like a Wi-Fi transceiver is sitting there just listening at one frequency for one sine wave and decoding that. No, no. This is happening at 52 different points all the time. And whenever we receive a transmission, we are receiving a BPSK, a 16 QAM 52 times across that 20 megahertz channel all at their own respective frequency spaces.

To me, it is absolutely mind blowing that this works at all. And we haven’t even talked about protocol level stuff yet. We’re just talking about the fire right now. It’s crazy to me that this works at all. And every time my phone gets kicked off of a network, every time my phone doesn’t work, I used to be a lot more grouchy about it than I am after learning about this and understanding how complex this all is and how incredible it is that 802.11 transceivers work as well as they do. I think it’s absolutely, absolutely fascinating.

OK, cool. So let’s move on. Thanks for letting me geek out on that stuff a little bit. What I want to talk about now is I want to try to demystify– no pun intended there by the way. I want to try to demystify spatial streams a little bit. Now, I’m not going to give you the absolutely perfectly correct engineering answer because to be honest, I don’t think I know that answer.

I didn’t go to school for electrical engineering. I don’t have a professional background in high level RF stuff. I don’t even have my ham technician license yet, which I need to get on that. I need to get my technician license. And that will happen at some point.

But I want to talk about the difference between a legacy single input single output device and a– we don’t have drawing again. There we go– a legacy single-input, single-output device and a modern MIMO, Multiple-Input, Multiple-Output device. And I want to talk about how that applies to missed access points and how it’s important to understand this when you’re speaking things like external antennas, things like that.

Somebody came off the mute button again. So I’m going to go find them and take care of that. All right, got you. There we go cool. Thank you. You beat me to it, sweet.

So let’s talk about what a single-input, single-output system is. Now, you’re going to see this graph. You’re going to see this diagram a lot in the world of RF. This is a simplified view of– we have a transceiver right here, and then we have an antenna. So that’s what we’re looking at, a radio transceiver and an antenna.

So with a single-input, single-output system, we have one radio and one antenna. That is a SISO system. Now, we didn’t call it SISO back in the day that I know of. I think we only– it’s a bacronym that we applied to this after MIMO came out, started to become a thing around 2007 or so at least in the world of in the world of Wi-Fi.

So what we do with a SISO system is we transmit that sine wave, and the device on the other end receives it, demodulates it, and there we go. We just talked about how that works, how that demodulation works. But the problem with a SISO system is that it is susceptible to multipath. Multipath occurs when we have reflections.

So, let’s say, for example, that when this transmits, one of those sine waves goes directly to our receiver on the other end. But one of those also bounces off a few walls and arrives from a different direction. One of them bounces around like this, and then eventually– I don’t know why it does that. One of them bounces around like this, and then arrives at our receiver on the other end. And what we get when this happens, this is multipath. These are the echoes that I was talking about earlier.

When you are in a room with a lot of reflective surfaces, and you hear your own voice echoing around, that’s because your voice leaves your face, bounces off those surfaces, and arrives back at your ears all at slightly different times. And what happens with that is that those signals, when they arrive at our receiver, they all arrive slightly out of phase with each other.

Now, we’ve talked about phases. But when a bunch of signals arrive slightly out of phase with each other, it starts to look like this. That’s what it looks like for the receiver on the other end. And for a radio receiver, you and I can– we can see several waves here. But to a radio receiver on the other end, it’s going to look like garbage. It’s not going to be something that it’s going to be able to demodulate. And so this was a major, major problem for Wi-Fi.

I remember back when this was my router at home, there was one spot in my house– you guys remember these. The good, old Lynx WRT54G. We like to laugh at these now, but I think that these are personally responsible for popularizing Wi-Fi in the home. And I think that ultimately, it plays a huge part in why enterprise Wi-Fi is such a big deal today. So good job, little buddy. Thank you for your service.

What this device was susceptible to was multipath. There was one spot in one of the houses that we used to live in where if I sat in bed with my phone, if I had my– and I just got in a smartphone at the time. I think it was a T-Mobile G1 way back in the day. There was one particular spot. It was about 2 square feet that if I sat there, my phone would not work. Wi-Fi would not work on my phone. If I moved three feet this way or three feet this way, it would work just fine again. And I’m pretty sure that was due to multipath.

I’m pretty sure that, that in that one spot, the reflections caused all of those phases to be misaligned in such a way that my phone could not demodulate anything or the router could not demodulate anything, not sure which way it goes there. But one of them could not demodulate the other and thus 802.11 did not work. This is a big problem. This was a major problem for 802.11 networks. And so what do we do to fix these?

Well, one thing that we could do and one thing that helped a little bit was single-input, single-output with diversity. So what diversity gives us is it gives us one transceiver and two antennas with a little switch in the middle. And so we can toggle back and forth between which antenna we’re using. In fact, this WRT54G has antenna diversity. So we’ve got a left diversity and a right diversity.

Now, I’m thinking back to when these first came out and when I was helping set some of them up for my friends. One of my friends got DSL. He had a 246 kilobits per second DSL. Holy cow, that was amazing. I was used to 10 kilobits per second of throughput on my dial up connection at home. And we got him one of these. And we had it in one corner of the house.

And I remember looking at the configuration utility for this and seeing that you could turn on and off left and right diversity. And I thought, oh, well, it’s on this– that’s on this end of the house. We don’t need any coverage to the left, so I’ll just turn that off. And so I turned off left diversity on this, essentially cutting off one of this poor WRT54G’s ears, so it could not use the other ear to try to better demodulate these signals.

But with a SISO system with diversity, what we’re doing is we can switch back and forth from antenna to antenna and try to find the antenna that has the best reception with the client that we’re trying to transmit to or hear from. That’s all we can do. We can switch back and forth between antennas, but we can’t look at both of them at the same time. So that helped. That could help with coverage problems, things like that, just a little bit. But it’s not going to help us with that reflection problem. It’s not going to help us with that. So what do we do?

Well, that’s when Multiple-Input or Multiple-Output or MIMO came along. What MIMO gives us, MIMO gives us one radio, one radio chipset with multiple transceivers built into that chipset. So in this example, I’m showing a 3 by 3 chipset. So this is three spatial streams, three transmit, three receive with three antennas. So we get three– we call these– each one of these we call it a radio chain, three radio chains.

Now, I think we talked about this a little bit in part 1 or part 2. But we will get a different amount of spatial streams depending on a few different factors on the device. First off is size. We need to be able to keep these antennas a certain distance apart for there to be any advantage. So size is a factor when it comes to how many spatial streams we can cram into a device.

Another thing is battery life. More spatial streams, more transceivers, more radio chains, that’s more juice, right. And so it affects battery life as well. Finally is cost. This device, my Nokia 6.1, this thing was $180 on Amazon unlocked. It is incredibly inexpensive. And for the price, it’s been great, but it is a single spatial stream. It does not have multiple radio chains. Whereas my MacBook Pro provided by Juniper Networks is a 3 by 3 device. It has three spatial streams, three transmit, three receive.

My iPhone SE, which I hope to get next Friday and delve into the world of iOS, is a 2 by 2 device. They managed to pack two spatial streams into that. That’s becoming relatively common on phones now. Most enterprise grade access points, low end is usually two by two. That’s typically what we see at a smaller access point. And then high end, like our flagship EP43, is 4 by 4. So it has four radio chains and so four transmit, four receive.

So what does this give us? With a multiple-input, multiple-output system, what we’re able to do is we are able– see me erase this really quick. Somebody just came off mute. I’m going to go and clean that up. There we go.

So with multiple spatial streams, this essentially gives our radios more ears to hear with and more mouths to transmit with. So I’m going to explain this in a very simplified way. This is not going to be super in-depth and possibly not super accurate, but it’s an understandable way to help you understand the benefits of MIMO in a short and easy to digest way.

What we can do with multiple-input, multiple-output is that when we transmit, the receiving device on the other side can compare reflections from multiple radios. It has multiple ears to hear with, and so it can compare those sine waves, and it can take an out-of-phase signal, and it can line them back up again. It can put them back together. And that gives us what’s called maximal ratio combining.

When we take two sine waves that are out of phase with each other, when we line them up, the result is increased amplitude. We see a heightened amplitude from those two waves because we have now combined them back together. And that is called MRC or Maximal Ratio Combining. I think I actually had a slide on it here. Yeah, yeah, maximal ratio combining. So there’s the name right there just so you can see what that looks like.

The other thing that we can do with it is called spatial multiplexing. With spatial multiplexing, we’re able to put a different transmission on each one of those streams. Basically, we’re able to put a different stream of data on each one of those reflections. Since we have multiple ears to transmit with and multiple ears to listen with, we can get a faster data rate out of this by putting different data on each one of those streams, really, really interesting stuff. And that ultimately is what makes MIMO so great.

To me, the huge advantage of MIMO is maximal ratio combining. The fact that we take something that used to be a major problem– multipath, it used to be a huge issue for Wi-Fi– and now we take it and we use it to our advantage so that we can realign those out of phase signals and get more reliable transmissions, really, really cool.
Basics 3.5
– So let’s tie that into antennas now. I’m not going to go through all this again. I think we spent plenty of time on– we spent plenty of time on this last time. And I think we need to take a break really, really soon here anyway.

But last time, we talked about antenna patterns a little bit because I wanted you to be familiar with the concept of external antennas. Now keep in mind, just remember that every antenna out there has a unique antenna pattern. Every single antenna out there does.

The antenna pattern of this WRT54G is going to be wildly different from the antenna pattern for a Mist AP41 or AP43. They’re going to be completely, completely different, which is why it’s very, very important, when you do something like predictive modeling in a tool like [INAUDIBLE], always, always, always, always use the antenna patterns provided.

There is a generic AP option there. Do not use that. Do not use the generic AP because it represents a perfect antenna and makes a perfect circle. That does not exist. There is no AP that looks like that. And so no matter what you do, if you use the generic AP in there, your design will be wrong. And you don’t want to do a design that is completely wrong.

Just as a really quick review. Remember that when we look at an antenna chart, we get a horizontal plane and a vertical plane. The horizontal plane is looking down on the antenna, like looking down on a floor plan. This is sometimes called an azimuth view or think of it as a top-down view.

Next is the vertical plane. This is looking at the antenna from the side, like looking at floorplans for a house from the side to see what is the front and the back of the house. What are they going to look like? What are the garage doors going to look like? That’s what we’re looking at here.

So for example, if we were looking at an antenna here, a dipole antenna would look something like this. So here is our dipole antenna. I’m going to use my Trusty WRT54G again. If we are looking at the antenna from the side, that is our vertical plane.

I like to call it the elevation view. That’s my favorite thing to call this because that’s easy to tie together with a house floorplan, elevation view, like an elevation floorplan on a house. Yeah, that’s which one it is. I never mix it up when I use elevation and azimuth.

Azimuth is if we’re looking at it top-down. So here with my webcam here, that’s the view from the top. That’s our top-down view on that antenna. So that’s how we read– that’s the basics, the real quick basics of how we read an antenna pattern.

You remember that we talked about different types of antennas like omnidirectional antennas. The antennas that are here on my WRT54G. I’m getting a lot of mileage out of this thing today, by the way. More use, and it’s had in years right now. This is an omnidirectional antenna.

Now, it’s not a perfect sphere. The only transmitter in our solar system with a perfect spherical antenna is the sun. Because this has to have some structure in place, a dipole antenna can never create a perfect sphere. But it will create a mostly omnidirectional signature. It looks like a torus, looks like a donut. Imagine putting a donut on this antenna. That is the approximate pattern that we get from most dipole omnidirectional antennas.

So if you’re doing something like using a Mist AP43, you can attach an external omnidirectional antenna to that. There are cases where you might want to do that. Maybe you want to tuck the AP inside the building but have the antenna on the outside of the building. There’s some things like that that you can do. So just keep that in mind. There’s a lot of different cases here. I’m trying to open your mind here a little bit to cases when you might want to use an external antenna.

Another type that we talked about is a directional patch panel antenna. Whoops. A directional patch panel antenna will radiate RF in a specific direction. So if you want to cover maybe a parking lot, something like that. What if you want to keep your coverage cells as small as possible to limit your contention domains. You might use directional patch panel antennas in a warehouse environment.

Point them down the aisles. Maybe in a hockey arena, maybe you put patch panel antennas up in the– what’s the giant screen thing in the middle called? Something tron. The jumbotron, right? You might put a bunch of directional patch panel antennas in the jumbotron all pointed out at different spots in the crowd.

There are tons of great use cases for directional antennas. Be very, very careful to not fall into the trap of always just using the internal antennas on the AP. Always consider whether it would be a good idea to move to an external antenna or not. There’s some really great use cases for these.

Here’s an example of a high-density patch panel antenna. This one has a very narrow beam width. The beam width here is– I’m going to just eyeball it. I’m going to say it’s about a 30 degree beam width. That’s about what we’re looking at here.

But there’s another aspect of this antenna pattern that I want you to notice here. Notice that when we look at the pattern, we don’t just have one line here. There’s not just a single line that defines the pattern of the antenna, what the coverage is going to look like from that antenna.

Notice that there’s multiple patterns here. We’ve got a black one, there’s a red one, there’s a green one, there’s a yellow one. There’s all kinds of different patterns here. Well, what’s going on? Well, the reason why we see multiple patterns here is because this antenna has multiple elements inside.

Remember that for MIMO to work, we have to have different radio chains. We have to have different radios. Yeah, that’s taken care of by the chipset in the AP or in the client device. But in this case, we’re talking about APs. And then we have to have separate antennas.

OK. How does that work with an external high-density patch panel antenna or any other type of antenna like this? Well, what we do is that each one of these– each one of these antennas has multiple leads coming off of it. So we get multiple leads that come out of it. In this case, this one has six leads.

So we get six leads, and each one of them has a little connector on the end. I think the Mist AP43E uses RP-SMA, if I remember correctly. I think that’s the one that does. Maybe Jean or something–

JEAN: Yeah, you’re good. You’re correct.

– It’s RP-SMA? I know there’s one of ours that uses an N connector. I couldn’t remember which.

JEAN: It’s the outdoor AP.

– It’s outdoor one. Yeah, that makes sense. Beefier. N connectors are nice and beefy, nice big beefy connectors, so it makes sense for an outdoor environment. But what we’ve got inside this antenna is there’s actually multiple– it’s not just one antenna element. There’s actually six unique antenna elements built into this antenna.

And so that gives us a little diversity, a little bit of difference between all of those antennas so that MIMO can still function properly. And so that’s what you’re looking at. When you see a shape like this, you’re looking at all the different elements inside the antenna and how they compare to each other to help you understand, OK, how close do these all align to each other? We don’t want them to align perfectly. We want to align somewhat so that we get a little bit of diversity for MIMO to work properly.

So let’s talk about how to use that with an AP43, specifically. This doesn’t just apply to Mist stuff. This applies to everybody, but there’s a specific use case here that I want to talk about. So what we’re going to talk about is we’re going to talk about the bad, the wrong way to do it. And then we’re going to talk about the good, the correct way to do it.

This is from a website called badfi.com. I meant to put a thing down here, but I forgot to. Badfi.com. It’s run by my good friend Eddie Forero. I’m sure a bunch of you on the call know who Eddie is. Basically, this is a blog of bad Wi-Fi installations.

So as a wireless network engineer, your goal in life is to never, ever, ever end up on badfi.com. Do not end up on badfi.com. I’m very happy to say that I’ve never ended up there. I’ve done a few things similar to what I’ve seen on badfi.com. I’m not proud of that, but you don’t want to end up on this website at all to be ridiculed by wireless network engineers everywhere in training classes for until the end of time. You don’t want that to happen.

So what we’re looking at here is we’ve got– Jean, do you remember what kind of Cisco this is? Maybe you remember. I can’t identify these like some of you guys can. Is that a 1,600, or do you remember?

JEAN: I can’t– man, I am drawing a blank. It’s the first NAP they came out with is the modular one. [INAUDIBLE] part, it had the– I just can’t remember the part number off the top of my head.

BEN: It looks like the 1252.

– 1252. Nice. Was that Garth?

BEN: That was Ben.

– That was Ben.

JEAN: Thanks, Ben.

– Thanks, Ben. [LAUGHS] so this is an example of how not to handle MIMO. And so what we’re looking at here with this access point– and I love talking about this example. Whenever I get to talk about antennas, I love bringing this one up because this is as wrong as you could possibly get. I’m sure anyone that hadn’t seen this before is like, what is going on there?

What we’re looking at is this is a 3 by 3 access point. So it has three spatial streams. And also, it’s a dual band access point. And what Cisco did with this access point, nothing wrong with how they did these, is three antenna connectors are probably 2.4 gigahertz or 5 gigahertz, I’m not sure which, but it’s either 2.4 or 5, so there’s our three antennas for the 2.4 gigahertz radio. And then over here on the other side, we have our three antenna connectors for our 5 gigahertz radio.

Now on the Mist AP43, we do things a little bit differently. We have an antenna connector on there. That one connector is dual band. It can carry both 2.4 and 5 gigahertz in the same connector. No problem there at all. Absolutely no problem with that. It’s just two different ways of doing things. No problem either way.

I do like having dual band connectors because it makes things a little bit simpler. Since the AP43 is 4 by 4, that means there’s a total of six connectors on this thing. If 2.4 and 5 gigahertz we’re all on separate connectors, well, that would be– let me do the math here really quick– that would be 12 connectors on one AP. It’d be a little bit out of hand.

So with this Cisco, what somebody did is– I’m going to say this is 2.4 down here. I’m going to change my mind and say that’s probably 2.4. What they did is they took one radio chain and point to that and put it on its own single element antenna pointing in one direction. They took another radio chain and pointed it in a different direction. They took another radio chain and pointed it in a different direction. And so they have completely broken MIMO.

I don’t know if this is going to work at all. It probably does very basically work, but I’m sure it did not work well. And then they just left their 5 gigahertz on dipole antennas covering up there and down here but not away from it. This is a mess. This is a mess across the board. This is really, really bad.

And I’m sure that whoever did this had great intentions. I’m sure they didn’t mean to make a horrible mistake here. They just didn’t know what to do. And so I’m here to help you not make the same mistake.

When you get an external antenna to use with something like an AP43, what you want to do is there’s going to be six leads coming off of this antenna. And so on the AP43, it has a 4 by 4 radio, four transmit, four receive, and that’s in both the 2.4 and the 5 gigahertz band.

Those are all combined down into four connectors. So you have a connector that does one chain for 5 gigahertz and 2.4. Another that does one chain for 5 gigahertz and 2.4. Another chain for 5 gigahertz and 2.4. You see what’s going on there.

Four of those leads are going to get plugged into, I think, it’s on this side. They’re going to get plugged into that side of the access point. But one thing that we also do on the AP43 is we also have a third radio in there, a dedicated 2 by 2 radio that’s there for doing things like seeing how many neighbors we can hear. It’s there to do scanning for RRM.

It does a lot of different jobs, and it needs to also have some visibility into what’s going on in the world around it. Fortunately, this antenna is a six-element antenna, so it’s got two more leads that we could plug in there.

Now, you could– somebody might slap me for saying this. I guess it’s nice that I’m in my office right now and we’re doing social distancing, so nobody can slap me. But what you could do is you could use a four-element antenna. You could. And you could either leave those two disconnected or you could slap a couple of dipoles on them.

But why would you do that? I mean, really? At the end of the day, just go get yourself a six-element antenna. Acceltex makes them. Ventev makes them. I think there’s a few other vendors out there that make antennas that work with this AP. And so you can get an antenna with six elements, and get them all connected, and things will work well. You’ll have good performance from there.

Any questions about this before we take a break? I think maybe it’s time for a break. What do you guys think? Break time?

JOEL: Yeah. Hey, Joel, the one question that came out since we’re on the topic of antennas, someone asked, do you have a list of external antennas that have been tested and approved for use? I responded back. I gave them a link of the Acceltex link where they have antennas for the Mist AP.

But I don’t know. Do you have any thoughts on just making sure that EIRP and making sure that we don’t blow out those limits and things like that?

– Yeah. Yeah. So my understanding of exactly which antennas to use with the 43 and the 63, or 61 rather, are a little bit thin. My understanding– maybe Raj can comment here, too– is that we have approved two specific antennas for use with the 43. There’s two that we have verified that you’re not going to go above your EIRP. Your EIRP is basically your output from the antenna itself, right?

Because you set a certain transmit power, I’m just going to make up some numbers here, like 15 dBm, right? You have a certain amount of transmit power from the AP, then you have a certain amount of gain from the antenna.

And if you’re not careful, you can go above regulatory limits, right? So we’ve tested a couple of antennas with our APs. That doesn’t mean you can’t use others. You just need to make sure that you have a firm grasp of how to do these calculations and make sure that you’re staying within the legal limits.

Raj, you want to make comment?
Basics 3.6
– I’ll jump right in and talk about the different architectures that are out there. Hey, Jean, can you just give me a thumbs up or something if you see my screen and all that? Everything’s working?

JEAN: Yeah, you’re good.

– You know how there’s normally a little green square around the screen to tell you that it’s being shared? Yeah, it’s gone. I don’t know where it went. So I think RingCentral is losing its mind today. So traditionally, I would say that we had three primary architectures in wireless networking systems.

The first is what I would call autonomous APs. This isn’t really an architecture at all because what you get is a bunch of access points that don’t do any kind of coordination. They don’t have any kind of central management. They’re all running autonomously doing their own thing.

And so that presents a few challenges from an operations standpoint. Once you scale more than five or six APs, that gets really, really tough to deal with. For example, if you want to change an SSID, you’re configuring five different APs individually. Nobody does this anymore. Fortunately, this is going the way of the dodo. This is even going away in home networks now. Even home networks aren’t doing this anymore.

Now you can go buy a mesh system like Eero or Google Wi-Fi or Orbi or any of those systems. They actually have some kind of central management. And so from a traditional enterprise networking standpoint, yeah, I would still consider those autonomous APs. But they all have a centralized management system, which is really great.

The big thing I wanted to point out about autonomous APs is typically they receive their traffic. And they typically dump that traffic directly to the distribution system. We often call the distribution system, the DS for short in Wi-Fi. Shorthand for DS ethernet.

The wired side of the network is typically the DS. Now, you can have some scenarios like with a mesh system where the DS, the distribution system, can be over the wireless. But for all intents and purposes, DS means wired side of the network.

So when we take traffic and we dump it directly onto the distribution system, it goes directly to the switch. And the switch forwards that traffic, routers forward traffic to wherever that traffic needs to go directly. To compare that, we have to look at lightweight access points. Some people might call this a thin access point.

And my understanding is that Bob, who founded our company, he is largely responsible for creating this architecture. I believe Airspace– as far as I know, Airspace was the first company to ever do this. They basically invented this architecture. . And for years, this was great.

So the idea of a lightweight access point is that the AP itself does not possess a lot of smarts. It’s not very smart. It’s just one step above being a radio on the end of a cable. And what the AP will do is whenever it receives traffic, it will tunnel that traffic all the way back to the controller. Usually, this is a physical box on site. Then the controller dumps the traffic out on the distribution system from there. Almost always, these will all tunnel back to a central controller.

Now, some advantages of this is that first off, it’s centrally managed. Of course, that’s a great thing. When you’re dealing with anything more than just a handful of APs having some kind of central management is absolutely critical. All the access points can all be coordinated.

All the traffic gets tunneled back to the controller, which can make things easy from a network design and routing standpoint. Maybe that allows you to get a guest network out in front of the firewall, things like that. It can make things really easy from that standpoint.

So there are some advantages to there. But this is what we at most would call old and outdated. This to us represents something that is now as Wi-Fi has scaled to become something way bigger than it was when controller architectures were first invented. This becomes very, very difficult to scale for modern networks because now you have to have controllers for your controllers for your controllers for your controllers.

What if this AP dies? And so if that AP goes away and– let’s say that this AP is now dead. If that AP dies and we replace it with a new model, what if the vendor doesn’t offer this particular model anymore? OK, great. Now you have to upgrade to their new access point.

But that new access point has a code version that’s incompatible with your controller. So now you have to upgrade your controller. But that means you have to upgrade all of your APs because your APs don’t work with the new controller. Oh, it just turns into a giant nightmare. Just a huge mess. So this was great first time. It’s outdated. And things need to change.

And that’s when we get to what I would call a cloud AP. Now, cloud-managed access points have been around for a long time. Lots of vendors have been doing this for a very long time. Of course, Aerohive did this for years. Meraki did this for years.

A cloud-managed access point isn’t necessarily new. But there are many aspects of Mist, about Mist, and what we do in the cloud that makes our technology completely different than other cloud vendors that are out there. But before we get into that, I just want to take a minute and talk about what is a traditional cloud access point.

And for the things that I’m discussing here, Mist clearly falls into this category. So of course, it’s managed in the cloud. APs are all coordinated to the cloud. But they function a lot more like an autonomous AP from a traffic routing standpoint. Traffic is usually sent directly to the distribution system.

So in traffic, if we get a frame in on the ethernet side, we dump it right onto the DS. And it gets routed around on the network immediately. We don’t tunnel it back to– we don’t tunnel it back to anything, unless you use a Mist Edge.

We have a special hardware appliance called a Mist Edge that you can either run in a VM, or you can run a physical box on your network. And the Mist Edge functions as basically a tunnel traffic terminator, an aggregator for all of your traffic from all of your APs.

The use case behind this is what if you’ve got a customer? What if you have somebody that has a traditional controller-based network, they’ve got a Cisco, they’ve got an Aruba, they’ve got one of these networks, and right now with their current architecture, they are tunneling everything back to the controller?

A Mist Edge can simplify things for them a lot because now instead of having to completely rearchitect your network and get access all the way out to the switches, now we can just tunnel all that traffic back to the Mist Edge and let the Mist Edge dump that out on the distribution system.

This is purely, purely an upgrade simplification tool. Well, this is one of the things you can use it for. There’s been some new interesting use cases around working from home. We’ve come up with some very interesting use cases for this. But the original intention with the Mist Edge was a place to tunnel all of your traffic back to simplify an upgrade from a legacy vendor to something new like Mist.

I don’t know. Jean, other SCs on the call? Anybody have any more comments you want to make on that? Or do you think I got the gist there? I should have warned you a little bit before I said, hey, open mic.

JEAN: Don’t see if there’s any questions. I’m good.

PARTICIPANT: Just one quick question that if the controllers work on layer two only, do we need one controller per subnet?

– Jean, you want to take that one?

JEAN: If they work on layer two only, no, you do not. What you can do is that you could actually put the Mist Edges. And just make sure we’re not calling a wireless controller–

– I was going to say–

JEAN: –the services architecture. And they’re cloud-managed and all the great things that come along with it. But you don’t need one per subnet. What you can do is that you can cluster them in a central environment that could be in the data center. They could be where your firewalls are, for example.

A use case example is maybe you want to tunnel your guest traffic to the Firewall Edge, for example. But to answer your question is that, no, you don’t need one per subnet because what you can do is that you can tunnel the traffic from the AP to the Mist Edge.

And then the Mist Edge can have a trunk port connection to the infrastructure. And then there, you can actually put traffic destined to different VLANs, different subnets based on what you configure on your SSIDs and policies.

SIMON: Hey, Jean. This is Simon. Just a follow-on. So if I have firewalls in between my APs and my Mist Edge, what ports do I need to open up?

JEAN: It’s a port. It’s–

SIMON: It’s all right if you can text it back at the chat.

JEAN: I don’t remember off the top of my head. Raj, do you remember top your head what ports we use for the communications between the AP and the Mist Edge?

RAJ: This is a cobweb tunnel between the–

JEAN: No, it’s not. It’s L2TPv3 is the tunnel mechanism. I just don’t remember off the top of my head what port it is.

– Rob says 1701.

RAJ: 1701? All right, cool.

JEAN: 1701, there we go. Cool.

– He probably just had that. He probably just recalled that from the back of his head. Probably didn’t have to Google it.

JEAN: Like I can’t remember.

– I already had to memorize the OFDM data rates. I’m out, right?

JEAN: Yeah.

– Any other questions, comments about the Mist Edge? One thing I will say if you’re queuing up a question here is the important thing to keep in mind about a Mist Edge is I know I use the same icon here as a controller. It’s not a controller. That’s an important thing to keep in mind. It is an endpoint for L2TPv3 tunnel traffic. That is the purpose of the Mist Edge.

And another thing that I think a good differentiator to make here, just a kind of– Rob says that 1701, that is for the AP to build the tunnel to the Mist Edge. That’s what that port is for. I’m sure we’ve got documentation for that on our website. If you go to mist.com/documentation, you should be able to find all the information that you need there. If it’s missing for some reason or you can’t find it, definitely let us know. We would be more than happy to help you out with that.

So I think some important things to keep in mind, differences between legacy cloud vendors and Mist is that what a lot of legacy cloud vendors have done is they have taken an on-prem controller, and they’ve put it in the cloud. And because of that, they are bound to the limitations thereof.

What Mist does is we use a microservices architecture that can scale elastically. For example, other vendors– I don’t want to make this an us versus them and get all angry about it or anything. But other vendors, when they take a controller and they put it in their data center, it’s still a controller. It’s just not on-prem.

It’s still one box that’s limited by its hardware limitations. It can’t see all of the other networks that are out there. It can’t understand what other networks are doing. Mist is built on a microservices architecture with AI in mind.

And so what we’re able to do with that is we’re able to do tons of machine learning, tons of data gathering to get a ton of data about what is going on with networks all over the place and use that data to improve all the networks that are in our cloud. So very, very different from other vendors and how they do cloud. And so I just want you to keep that in mind. It’s a very different ball game here. OK. So with that–
Basics 3.7
JEAN: Oh, wait a minute. I got to at least start my video if I’m going to talk a little bit.

– No worries.

– Yeah, but anyway, so let’s talk about three things. And I think we reviewed these before. So if anybody has any questions, go ahead and let’s feel free to shout them out, put them in the chat. Let’s make sure that everybody is comfortable and confident in how to use some of this stuff because it’s like key. It’s a foundational thing when you’re doing a pilot or a POC, or testing in your lab, or things like that in terms of some of these things that I just want to quickly go through.

One, is Joel asked me to kind of go through config templates. And this is live demo. And one of the things everyone is that everybody should have access to live demo, which is really cool, go in here, demo, and do all the things you need to do. But please, please don’t make any changes.

So just be very, very careful of that because live demo is actual– it’s an actual live demo network that’s running. People are connecting in the office when they go there. There’s not many people there right now. But when they go there, they are using this from a production perspective.

So let’s start up with config templates. And you can see in here in live demo, there’s a whole bunch of stuff in here. And Raj, we probably want to clean some of this stuff up.

Looks like there’s some really interesting things in here that probably aren’t being used. But we should probably make it look better because it kind of–

RAJ: No, it keeps coming, Jean.

– So everybody, please, please let’s not make any changes because this look does look a little strange. So– but the first thing–

JOEL: At least Vidur put his name on his right here, right?

JEAN: Yeah, hey, we know who did this one.

[LAUGHTER]

Yes, but so, just really quick, a few things to look out for. Like, I’ll go in, and we’ll look at what configuration template looks like really quick. But the cool thing is that, look, at first glance, what you can do is to see like hey, how many templates do I have applied, right? So you can create templates.

And just kind of as a rule of thumb, what I like to do is when I’m talking to a customer is say, look, what SSIDs, what security policies like that do you have that may be across your entire organization, right? Maybe they have one campus. Maybe they have a campus with 500 sites.

And they want the same SSID, for example, everywhere. And so I kind of start by getting an understanding of what they want to do and then that actually helps guide me along in terms of the templates to configure and then how I want to apply them, right? So you see here. Here’s the name of a template.

And we’ll go into one of these here and take a look at it. And then, you can see here where is it applied to? Is this applied to my entire organization? Or do I just have a template that I want to apply to a few sites, maybe one site.

Maybe you want to create a template for your lab testing, right? Maybe they just want to go do some testing. So you can actually do that too, right? So there’s a lot of flexibility here.

And then obviously over here, it says, hey, if you don’t apply it to the entire org, what sites is this applied to? And then there’s some stuff in here with site groups that you can actually have a little bit more granularity and flexibility and then some exceptions. We’ll show you what that is.

And then here’s the WLANs that are actually applied to this template. And one of the things that I recommend that we’ve talked about on these calls– Joel’s covered this before– is hey, for SSIDs, try to minimize how many SSIDs you have out there for lots of different reasons.

But– so let’s click on one of these. And when you go into these, you can see here this is kind of a template that someone created called test one. There’s nothing in it. And then all you need to do is there’s some stuff in here for tunnel configurations and policy stuff. But I’m just going to focus here on WLAN.

So what you can do is you can click on Add a WLAN. And I can call this ABC. And I’m going to do something that I told everybody not to do.

I’m just going to do a quick configuration here. So you guys can see what it looks like. And the risk is low because I know– and I’m going to go through this really quickly– is that this– if I do make a change to this template, I know it’s not applied to any sites. And then I know that there’s not any access points that are also bound to this thing.

So I’m not going to go break something right now. But I can create an SSID called ABC. Let’s just create a PSK. And then, of course, when we go through here, here’s where you would actually configure your data rates that you want for your environment.

And you can do custom rates. You can do the– have all the data rates turned on, or you can do something in between. We actually have this documented in terms of what obviously no legacy means, no G, or B. But the high density rates– I can’t remember. I think it’s– I can’t remember if it’s– I think it’s 18 megabits is the lowest data rate enabled on high density.

So if somebody can remember and shout it out that would be great. But anyway, you can figure this thing. Here’s where you actually can configure what VLAN do you want this assigned to. Do you want it to just go on the untagged native VLAN? Or, do you want to actually apply it to a tag VLAN, for example, right?

Once you create it, it’s going to pop up in here. And then once you have your WLANs that you want to create for your organization, now what you can do is that you can say, hey, I want to apply it to the entire org. And I’m not going to push that button because that’s something I don’t want to do, even though I’d have to save it, or, I want to apply this to certain sites or site groups that you can create. So that’s kind of how you do it.

So I think everybody is somewhat familiar with this. But I just wanted to quickly go through and show you this to see if anybody had any questions, and then that way we can knock them out now. And one of the things I wanted to show here is that if you’re playing in your labs or you’re at your customer site, and you go in here, you can see these templates you’re creating, here are the sites that it’s bound to. What I like to do is that I like to go to my inventory page. And I can see that I have all my access points.

You can see here that I got some connected, some disconnected. But the nice thing is that I can see over here these access points, what sites these access points are actually assigned to. And then that way when I make changes to these templates and I say I want to apply this template change to site live demo, here are all the access points that change is going to get applied to. Right now this is just in one site. But this could be hundreds or thousands of access points spread out all over the place.

So templates are fantastic. They’re powerful. But just be careful in terms of how they’re being used.

Now the one thing I do want to point out is that sometimes I see where templates aren’t being used. I’m a fan of templates. I like to use templates.

I like to build a foundation and get started there and kind of help customers also think about using templates to get started. So you can do some cool things. I’ll give you one example of what I did with templates.

I was working with Petco. And Petco– we had our we had our SSIDs that we had that we wanted to deploy everywhere in all the stores. However, they said, look, we’re going to be upgrading some of the stores. We’re going to be adding guest access.

But we don’t want to do this to all the stores because they have T1s now. And if we turn on guest access, they’re going to blow out our WAN circuits. But they were upgrading some of their stores from T1 to a broadband. And they were using SD-WAN and things like that.

So what we did was is that we created another template that was for guest network only. And then now what happened was is that every time they upgraded a store, and they increased– they upgraded from T1 to a broadband connection, they can go add that store to the guest network template, right, now that we’re enabling guest networking as they were moving forward and migrating. So you could actually have two different templates that actually are applied to the same site, for example. That’s just one example of what we did there.

The other thing that I just want to make sure that everybody is careful of here is that you could create a WLAN. For example, you see here I got this one called SSID ABC. You could go into network and under WLANs, you can see here that this is actually really clean.

You see that here are all my SSIDs. And I know that all the SSIDs that I have running in my production environment on live demo are all bound to my template. Fantastic. That’s what I like to see.

However, I could click on WLAN and actually add a WLAN here manually without using a template. I can call this SSID ABC.

And now I could potentially have two SSIDs called ABC, running on live demo. We don’t want that, right? So it’s some interesting flexibility we have but something to watch out for.

Every once in a while, I see that where some folks get confused between templates and actually creating an SSID specifically for that site under the network setting. So that’s one of the things I just wanted to point out for config templates. If anybody has any questions, I don’t know if there’s anything in the chat, Joel, or if anybody has any questions, shout them out now before I turn this over to something else.

– Nothing specific about our F templates. Nope.

– Perfect. So let me go to RF templates and a couple of tricks that I like to look at here. One is that I know we’ve talked about this before from an RF template perspective.

This is– just from a baseline perspective, what I like to do, and what we recommend from a best practice’s perspective is to go in and create your templates and create your templates for both 2.4 gigahertz and five gigahertz.

Man, if I had my way, I would turn off 2.4 for every customer. But obviously, we can’t do that for everybody.

JOEL: Preach it, brother.

JEAN: What’s that, Joel?

– I said preach it, brother.

– If I could. However, I understand that there’s 2.4 that’s out there and whatnot. So a lot by the way.

And so what we do from a recommendation perspective is to go in and actually tune your mid max power settings for 2.4 and five gigahertz. And you can see here in live demo, this is one thing I wanted to point out. You could point out is that you want to create some separation in terms of your transmit power between the two bands.

So that way when the clients hear the SSID on both 2.4 or five gigahertz, they’re going to choose five gigahertz. If they’re five gigahertz capable, they’re going to choose that because they hear that better. And it’s really important to get this tuned appropriately to do that.

And again, just from a recommendation perspective, these are some guidelines that we go on. You want to create a few DB of separation between these. And when I say a few DB of separation, I’m going to kind of show you how you can kind of see that and what makes sense.

But typically, when I set something up, I’ll default to something similar to this just to get things up and running. And then I’ll take a look at what’s going on in the environment and then potentially make a little bit more tuning if I need to. And let me show you two things that I kind of look at and at least two things that I look at to determine that.

One is once I do this and once it’s up and running, what I do is I go to the WIFI clients page. And I want to get a good idea. You can see here, I go, look, I’ve got 22 clients connected right now in live demo. And over here on– oh, let me take that back. I already put a filter in.

I have 25 clients connected on live demo. But how many of those are connected to five gigahertz, right? Yeah, here I can easily eyeball that thing. But if you put the filter in, five gigahertz, now I can say out of the 25 clients, how many are in five gigahertz? Wow, 22. That’s pretty good, right?

I’m really happy with that, that I have 22 out of 25 clients on five gigahertz. That’s telling me that I have my RF template configured pretty good because now most of these clients that are connecting are preferring to use the five gigahertz band. So that’s one thing that I go look at just to eyeball some things.

The other thing I look at is that I go under network and radio management, and I look at the 2.4 and the five gigahertz bands. And I try to get an idea of, hey, what’s my– what are my channel configurations look like? You can kind of hover over these and see how many AP’s are on each of these channels.

Same thing for five gigahertz. And then– but the other thing that I look at with regards to my RF templates is I look at the power. And I get an idea. You saw in there that I had 2.4 gigahertz set at five at the min and 7 at the max.

And for five gigahertz, it was 9 to 17. I think that’s what it was. So I can get an idea of what my power settings are across my AP’s just to see, hey, is this looking pretty good from not only from a channel perspective but also from a power perspective. And in here, I can click on 2.4 and go, hey, look at this.

This looks pretty good. I set all my 2.4 gigahertz AP’s and my RF template five DBM, and that’s what they’re coming up as. They’re not even going all the way up to seven. So that’s actually pretty good.

And for five gigahertz, the minimum was nine. Looks like I already got them between 10 and 18. If I were to look at this, I’d be like, hey, how come I have so many that are actually running at a power level of 18? I may want to take a look and see do I have enough five gigahertz coverage and things like that?

But I won’t dive into those details. But just I wanted to point out that, hey, here are some things to look at just to kind of validate do you have your RF configuration template, kind of dialed in at least close to where you want it to be to provide a great experience for those clients. Any questions out there, Joel, any comments from anybody?

JOEL: No, not around this. I got a couple of other questions that aren’t related. But I mean, I guess we’re kind of talking about operational stuff. So we might as well.

One question is a very basic question is, how do we handle AP software patching? Is the same capability available for EX switches? Or, do we only do it to telemetry for EX today?

– Today, we only do telemetry for EX. But obviously, once we introduce the configuration capabilities, we will obviously be able to manage the firmware operating system updates. So from an AP perspective– and by the way, when it comes to wired assurance, I don’t know for sure, but I believe we’re going to have a similar way to do that. And I’ll actually show that here in just a second.

But one of the things I look at, I look at access points. And over here, I can see already what version I have. And then of course, I could sort on these different versions.

And then I could actually do– I think I can actually filter on these things as well, right? So you can get an idea of what’s going on out there. However, what I use– and the reason why the question came up is that I really like to make sure that all my AP’s are on the same version, right, something that I know is stable and working and all that stuff.

So what you can do is you go to your site configurations, live demo. And over here is where you can actually control what you want to do from a software perspective. And so you could– there’s a couple of different ways to do the upgrades. But one of the things you do is that you could also could set all of this stuff up upfront before you even plug-in a single access point. And you know that I want to make sure that all my AP’s are on firmware RC1. And then you set that stuff up here, enable AutoUpgrade, enable RC1.

You can go select, or you can go select a custom version if maybe our support organization recommended or whatever. You set that up ahead of time. And now when you plug-in an AP, and you assign it to that site and it connects to the cloud, it’s actually going to get the firmware version that you specified here. So that way they’ll all be on the same version.

But– so I recommend that because it just happens automatically. It keeps everything in sync. And you don’t have to worry about AP’s coming onto the site with a different firmware version.

However, sometimes folks may want to do– maybe want to test something in their lab. Maybe let’s say one of these AP’s are in a lab environment and you wanted to test the latest and greatest firmware version that came out because some new features in there or fix or something like that, so you can take a single access point, click on it there, and then I can go to upgrade AP’s, click on proceed, and then select a version that you want to upgrade to. And you can actually do it to a single access point.

I could do the same thing to all of them here if I wanted to do the same thing. So there’s a couple of different ways to do that. But I like to control it at the site level. So that way when AP’s come in, they just get the right firmware version on it from the beginning. And you don’t have to worry about doing the stuff after the fact. OK.

– I really like that you went straight to the AP at Raj’s desk.

– Yeah.

– [INAUDIBLE] AP, that’s the one.

– That’s the one. Yes, any other questions? Because I’m going to talk a little bit about security before I turn it back over to you. But is there any other operational questions that we want to tackle that are there right now?

– Nothing’s come up.

– Perfect. So one of the things– this does come up a lot, right, security wise. And by the way, I use security for a lot of different things.

And so real quick, just talk through what does this mean. So we have rogue AP’s. And what the definition of a rogue AP is that of course, the missed AP’s can hear it within the site. Over here, you see here’s the site up over here.

The AP’s can hear it. But in this case, what it’s saying is that not only can they hear it, but they also know that this AP is connected to the same physical wired network at least in the same subnet as the missed AP’s. And how do they know that?

Well, I don’t see any here. But over here when clients are connecting to these things, what happens is that we hear the clients. I wish there was one– let me do this. Let me see. Is there any in the last 24 hours?

Aha, here’s one. See right here where I have a client connected to this? I can click on that. And now I go, look, hey, here’s a rogue AP.

I hear the client that’s talking to the rogue AP. Oh, by the way, I also hear this client on the wired network. So by definition, I know that this particular device is connected to the network because it hears this client that’s connected to the rogue AP.

It sees that Mac address or hears the Mac address on the wired side. That’s how we know it’s connected to the wired network. So that’s the definition of a rogue AP. And obviously, we have the clients connected to VSS ID, what channel it’s on, the signal strength at which we hear.

How many missed AP’s at the site actually here this particular rogue access point? And also, what is the nearest AP to this thing? And then if you have these AP’s that are on– if you have a floor plan and you have all the access points on the floor plan, what it will do is that you can click on the location. And it will tell you, hey, here’s the nearest missed AP.

And where that missed AP is located to where it actually hears that rogue AP, right? So it gives you an indicator of, hey, what area do you need to go to actually go find that thing.

– Hey, Jean, there’s a question. Do you need a specific firmware to see rogue AP’s?

– Oh, the answer is no. But that is a really good question and something that I totally forgot to actually cover. So thank you whoever asked that question, because let me show you guys what you need to do.

And I totally missed that. And I meant to talk about it. If you go to site configuration, I’m going to go to live demo. By default, the– where is it?

– It was on the left hand. Scroll down a little bit. I think I know what you’re after. Right here.

– Where did I miss it?

– Security configuration.

– Oh, gosh, that’s right in front of me, Joel. Thank you so much.

[INTERPOSING VOICES]

– Happens to me all the time.

– You are master of those annotations, Joel. See, look at this.

– As long as I can annotate on your screen and not mine [INAUDIBLE].

– Thanks, man. By default, boy, it was right in front of me. I didn’t see it.

So by default, that’s turned off. And so by default, when you create a site, go in there, turn that thing on. And then what you can do is is that you can say, hey, I don’t want to be able to detect rogue and neighbor APs. And then you can actually say, look, I want to be able to– I only want to document that or post it in there if I hear a rogue AP at a negative ADDB or better and then also, if I hear it for 10 minutes or longer.

And the reason why we do that is that a lot of times– Joel was talking about all the ways that MRC and all that stuff work with the signals bouncing around and all that stuff. Well, you know what? The same thing happens with this.

Our APIs here, rogue APIs, out there, signal may bounce around. We may hear it real briefly. And it may just go away, right?

It may not be something that’s a constant thing that we’re hearing. So that’s why we actually have this timer threshold in here as well. We want to make sure that we’re consistently hearing something. So when we report it that it’s pretty accurate in terms of what we’re hearing and and for the time duration.

So thank you for asking that question. Yes, you just need to turn it on. That’s all.

So let me go back over here. So hopefully that answered that question. So rogue AP’s.

Honeypots. So honeypots is very similar to rogue. But honeypots may be the fact that it may not be connected to the physical network. Or, maybe we need to hear it on the wired side.

However, it has the same SSID that we’ve configured on the MST AP’s for our production environment or whatnot. So that’s basically what a honeypot is, same thing over here with the channels, how many AP’s hear it where it’s located and all of that stuff.

Hey, neighbor AP’s are just, hey, what do we hear that’s around us? And how– and what do we hear that’s around us? And then how hot are we hearing things? So that’s what that is.

And then approved AP’s, these are the ones that are actually– so this list here is the same list over here that we have on underlie demo, right? So that’s all that’s saying.

Now, the one thing that I use this for is obviously to show customers what we do from a rogue perspective. I help them understand the definition honeypots and things like that. But boy, the one thing that I use this for is that when we’re working with customers and we’re doing a cut over from like mist to something else, and like, let’s say we’re replacing Cisco and they got a bunch of AP’s out there, and we’re replacing them, I got to tell you, nine times out of 10, they go, hey, we’re all good. We’re all done. All the other vendor stuff is gone.

But you know what? You go in here, and you go no, no, no, there’s still one. There’s still two, three AP’s floating around, we’ve got to find them because boy, if they’re sticking around out there, they’re going to provide a really bad experience because that client is going to go try to connect to these things. And they may not even know, or it may be that something’s out of commission or something like that. So anyway, this is a fantastic way to go find stuff that’s out there that the customer thinks they’re gone. Anyway, so that’s why I use it all the time.

– Garth is screaming Northgate in the chat.

– [LAUGHS] Yes. Yes. Yes. Yes. Yes.

Yes, Northgate was one of those that they said no, no, it’s all gone. You go in there and go, no, no, there’s still some around. We’ve got to go find these things.

Yeah, and can you imagine this? Like, at Northgate, they had AP’s like above ceiling tiles plugged in, right? So they didn’t even know where they were. But yet, we could hear them.

We helped them find them because we get the nearest AP. And we can help them figure it out. And we did it. But yeah, we– it’s very, very important to make sure that these little extra floaters out there are gone.

– I’ve heard stories of AP’s being, like, stuck in a wall, dry walled over, these type scenarios where there was literally no physical access, no way for a human to interact with that AP because time has locked it in the wall or something like that. So–

– Yes.

– Awesome.
Basics 3.8
– But I just want to talk about how RRM works. If you’re not familiar with RM as a concept, RRM stands for Radio Resource Management. And RRM allows the system to automatically tune things like transmit power and what channel it’s on. Those are the two big things that RRM will tune.

MIST has a kind of a unique way of doing RRM. We kind of split it out into two categories– global RGM and local RGM. So global RGM runs every single day at 3:00 AM local to the site. So whatever the local time is on the site, it’s going to run.

And it’s going to use the last several days’ worth of data to make informed decisions about what channels the access points should be on. It’ll figure out what the transmit power should look should look like based on neighboring access points.

It uses all this accumulated data and kind of functions like a reset button for the whole site. So it can make a massive change on the site all at once. And that’s why we do it at 3:00 AM.

I think one thing that’s coming fairly soon is that one of our critical customers or one of our customers that I would say is MIST is absolutely critical for them, and absolutely has to work, is a large shipping company.

Yeah. 3:00 AM is one of their peak times. And so before too long here, this will be configurable, so that you can set when you want global RRM to run. It can also be manually triggered.

So I don’t know if you noticed, but when Jean was going through the dashboard there, up at the top, you could actually hit– it’s kind of– I’ll just draw where it is up here. It’s kind of up here in the corner on that RRM view. And it says, optimize 2.4 gigahertz now or optimize 5 gigahertz now.

And that will run that reset button using that data set to reset what channels and transmit power everything is on. And this uses reinforcement learning over a long time to figure out what everything should be set at. And that’s great. That’s really nice.

But the problem with that is sometimes, there are events that happen during the day. Well, a lot of times, there’s a lot of events that happen during the day. And so we need to be able to react to small changes. And that’s where local RRM comes in.

So what local RGM does is it runs on an ad hoc basis whenever it’s needed. So it reacts to small events like, for example, let’s say, that you have an AP on channel 52, up in the 5 gigahertz band. And somebody outside in the parking lot fires up a quadcopter with a 5 gigahertz analog video transmitter for their first person view goggles or something like that.

Well, if we have something like that, something high power show up on our channel, that could cause a clear channel assessment back off to trigger. And so none of our clients will talk because there’s something talking on that channel. They might detect, oh, there’s RF there. They’re just going to stop. They’re going to back off. They’re not going to do anything.

OK. Well, what do we do? That channel is essentially being jammed. So what we need to do is we need to switch to a different channel. So what we’ll do is it will react to a local event it can do things like switch to a different channel to get away from interference.

It’s completely independent of the cloud. We don’t have to wait until 3:00 AM. This can happen at any time. And there’s lots of different events that can run like, for example, interference from another access point on the same channel, interference from a non Wi-Fi device. Maybe radar was detected on the channel, we need to move to a different channel. So there’s a lot of different things that can happen there.

So I think I will jump into the dashboard real quick. I’m going to try to make it as quick as I can here.

– Hey, Joel.

JOEL: Yeah.

– While you’re jumping into the dashboard, I just want to make a couple of comments on this. In this RRM overview that Joel just went through, I mean, this is what we do here is something that a lot of– that no one else can do, basically. And it’s super, super powerful.

When he talked about global RRM, think about the fact that Joel talked about the AP41, 43, how we have a dedicated scanning radio. We use that dedicated scanning radio to listen to what’s going on over the air, 24 hours a day, seven days a week, all the time.

And a lot of times, our customers say, hey, what do you have that’s in your system that’s AI-driven, and things like that? Well, you know what? This global RRM is absolutely one of our AI-driven mechanisms that we have built in.

It’s a deep learning AI. It’s a deep learning algorithm that’s ran because we’re collecting all that data all the time. And then he mentioned reinforce learning, right? So before we make changes and power and channels and stuff like that, we leverage reinforce learning to make sure that, hey, before we make these changes, we’re going to provide a better experience in this environment.

So some things like that, that we leverage the cloud and the microservices architecture that we built. So think about the power behind that in terms of all the APs and all the data that we’re collecting. That’s something that a controller solution cannot do. It just can’t do what we can do from an RRM perspective.

So definitely keep that in mind when you’re talking to customers about the power of how RRM works, and what we can do.

– Awesome. Thank you, Jean. Much appreciate it. Give me some time to get transitioned over here too. So Jean showed you how to get in here already, but you go to Network, and then you point to Radio Management and click on that. And that brings you to this RRM view.

I can’t draw on this. It’s going to drive me crazy because normally, I scribble all around on here. But the things I want you to look at for, notice that we can switch between 2.4 and 5 gigahertz. We can switch from whatever site that we want here.

So ours is called live demo. But usually, you’ll have a code or something there to say what site it’s for. And then here’s that optimized 2.4 or optimize 5 gigahertz now buttons up here. There’s a few indicators up here at the top to tell you how things are doing.

A big one that I always look at is the average noise floor on the site. One of our customers, they had a site where data rates were just awful, just across the board. Throughput was down, things were just not great.

We looked at the average noise on the site, it was like negative 75 dBm. That’s where the noise floor was. And that eats away at our SNR, which you know what that does. That means that we can’t do higher order modulation schemes. And so that starts to slow things down and eat up more time on the channel. And everything just kind of spirals out of control. So I love having this indicator here.

Average neighbor says what is the average amount of neighboring access points that we see per AP. And Jean, that includes everybody, right? That’s not just ours. Average neighbors is everyone, I think.

JEAN: Yeah.

– That includes other neighboring networks, right? Another one that I recommend looking at is the channel distribution score. So 0 is bad, 1 is good. That means all the APs are very evenly distributed between channels.

You don’t want a bunch of apps all piled up on one channel, so I like to keep an eye on that. And AP density says, how many of our APs can hear each other? We want that to be somewhere between, I think, I’ve heard somewhere between 0.7 and 0.8. It’s kind of where we want to see that.

Our density is actually pretty low on this site. That could indicate that maybe we have some coverage gaps something to watch out for. So Jean, pointed out the distribution here. This shows us how many apps are on which channel.

So for example, channel 128 has two APs on it. Channel 149 only has one. You can also look at power there. What is our power distribution? So we have one AP that’s at 13 dBm. We have three APs that are 18 dBm. So you can kind of get an idea, where’s my transmit power lining up here.

If you scroll down, these are the current radio values. What is everything configured to right now, here’s our number of clients, what channel we’re on. Those that we have our primary 20 megahertz channel. And then our secondary 20 megahertz channel, which equals a 40 megahertz channel in total.

There’s some great things that you can investigate in these little expanders here. Like this one, this is looking from LD Deckert Mesh Bas. It’s the AP that we’re looking at right now. We can see, using that dedicated third radio, we get some interesting information about channel occupancy in the area. What’s going on with the nearby channels?

And you can sort them by just channel number, or you can go best to worst. And predictably, channel 36 has way more going on than channel 161. Never a surprise to see that channel 36 is always slammed.

And the color tells us what is going on there, whether it’s activity from external access points, our site access points, or even non Wi-Fi activity will show up in red. So lots of cool stuff there. Go check it out. Just go look around in there and see what you can see. Yeah

– Sorry. One thing while you’re in there. I just wanted to point this thing out because sometimes, it gets overlooked. If you can just click back in there. You can see the channels that are– click if you–

JOEL: What happened? Oh, I got to scroll down.

– See how the numbers are bolded over there. So those are the channels [? we design. ?] Just wanted to point that out.

– So LD KRaj. This is the one next to Raj’s desk. I don’t know if it’s on his desk or above it. I like to think of it as being on his desk. I’m sure it’s not, but I like to think of it as that’s where it is. It’s on channel 149 And 153. So you see that right here as well. That’s a good thing to point out for sure. That’s up in you need to extend it. Where is that?

– Yeah. You don’t want to see my desk now.

JOEL: Nice.

– And one other thing Joel that I wanted to point out is sometimes, this gets overlooked. And it’s a really cool thing. See in the– and that window there, where it’s just current radio values–

JOEL: Yes.

– If you click on that little arrow, and I didn’t mean to steal your thunder.

– Oh, I’m really glad you brought this up because I want to talk about this, but you go ahead.

– Oh, no, no. I just wanted to make sure you pointed that out because you get a [? vague ?] view of [? what is ?] in the channels.

– Yeah. This is great. Thank you so much. I was about to skip right over it. I’d forgotten about it. So what we get here is this is our floor plan view. Now we have dragged these APs onto the floor plan using the live view right here.

Really easy to do. It’s exactly what you would expect. You drag and drop APs, they show up on the floor plan, you upload a floor plan, you set the scale, all that. But the big things that we can see here is first off, we get the color tells us which channel it’s on.

The gradient’s a little too, I don’t know. It’s hard to tell exactly what’s on what channel. But you can tell that like this AP is on a wildly different channel than this AP. That can be helpful. But what I like the most about this view is that the size of the circle tells us transmit power. What is the current transmit power.

Now that doesn’t tell you what the coverage cell looks like. Remember we talked about that. Coverage cells are not pretty circles. This is not calculating free space path loss or anything like that. What all we’re doing here is we’re just showing you what is the transmit power relative to each other.

You can see that this one, on channel 124, that AP has got a way higher transmit power than this AP here on channel 52. So that can be really nice look in there and see what’s going on. Can you click on these? Oh, heck, yeah. I didn’t even realize that you can.

Yeah. So if you click on LD server room, that will show you where it is. Power is 15 dBm. It’s at 40 megahertz right now. And it’s on channel 60.

So down below, the very bottom part of this view, is radio events, either over the last 24 hours or seven days. So we could see what events have occurred, what channel it was on, what channel it went to, what power changes were made. Like in this case, there were none, no power changes.

Another place that you can see this, though, is if you go to Monitor and go to Service levels, you get all these little triangles down here. Let’s go to seven days worth of data here. All these little triangles indicate changes that were made.

The green ones are RRM changes that were made, whether a periodic one or a site wide change. And then the blue ones, these are all configuration changes made by administrators. So for example, Vidur went in here and he added a virtual beacon called Welcome. So you can see that.

So we can see exactly– Oh, looks like Phil added a map called Test. So Phil, messing with our live demo. Look at that guy. Typical Phil right there.

One question is what’s best practice, global RRM or local RRM? Here’s the cool thing about that, you don’t have to choose. They actually are both important parts of RRM. They both work in tandem to deliver the best results.

So global RRM is kind of that OK, overarching network plan using a large data set. What is the overall best plan for this network to help our service level expectations, to help the client experience? And that happens over a long time. Whereas, local RRM is an ad hoc thing that only runs to deal with certain local events.

And so local RRM events feed into global RRM as well. So for example, if we see, regularly, interference on a specific channel, then that feeds into global RRM. And the global RRM will stop preferring that channel.

There is no picking between the two. These are just both two pieces of how RRM works. Excellent question.

Yeah. And if you hit that Reset button that I talked about, that global RRM button, it does reset the entire site. So I wouldn’t do that if you were serving clients, if you’re actively serving clients on the site. That’s the kind of thing that you want to run at night, you want to run when nobody’s around, I would think.

I’ve never tried running it during the day. I have no idea what would happen. I would assume some radios would go down.

– One comment on this, Joel. You talked about the Optimized now button, that’s in the upper right hand corner. And one of the things I wanted to say is that hey, when you guys are working with a customer, and you’re spinning up a pilot or POC and you’re getting it running, what I like to do is turn everything on.

And then I don’t want to wait for global RRM because that happens every 24 hours. And it does its thing. So what I like to do is give the scanning radios, give the APs about 20 minutes for them to kind of start hearing what the environment looks like.

And then make sure you have an RF template set up. So you have that all ready to go in terms of what you want that to look like. And then I go click on Optimize now for both 2.4 gigahertz and 5 gigahertz.

And then what it does is it brings those radios into alignment to the base configuration based on what they’re listening to for about 20 minutes. Don’t try to do it after five minutes. Give it some time to scan, and then it will actually do a pretty good job for you when you do that.

– Cool. Question– personal question. In my house, I use Google Wi-Fi, spread across three floors. I’m considering getting an AP43, but want to make sure my clan stays happy as well. That’s smart. What’s the best way to figure out placement and keep my current setup?

OK. So if you want to keep your Google Wi-Fi stuff, personally, I think those things weren’t great, by the way. I haven’t used Google Wi-Fi, specifically. I’ve used Eero. And man, they worked. They just work really well. I was impressed.

Make sure you have neg 67 dBm coverage everywhere. That’s like the biggest thing that I would be looking for if I’m kind of checking out my home network and seeing how things are going. Those systems seem to do a pretty good job of picking good channels and things like that.

If you want it to get an AP43, though, and I recommend that you do. Because if you want to really– if you want to dig into this stuff, man, there’s nothing like hands on experience. And getting an AP at your house is just a great way to do that.

You might want to do more than one. For example, I have two vendors here at home. Of course, I’ve got MIST and I’ve got another vendor as well. That way, if one of them goes down, I’ve got something to fall back on.

And I’ve personally caused problems on my own network all the time. And I want to keep the clan happy. But just because it’s an AP43, doesn’t mean it’s going to cover three times what one of your Google Wi-Fi devices would.

I would expect an AP43 to have a better reach than the Google Wi-Fi ones would. But I wouldn’t expect it to cover three of them. So if you have ethernet in the house, then that’s great. Get a couple of APs in there.

You could also consider doing– there’s a single hop mesh mode that MIST does. So you can have a root node, plugged into like an EX2300 in a closet somewhere. And then you could put a couple more APs around. Just make sure that they have very good signal strength back to your root node and you should be OK.

Question here. I need some help with this one, Jean. How will 5 gigahertz clients affect mesh networking on MIST mesh Wi-Fi?

– They just don’t share the uplink. The radios will use 5 gigahertz to connect to each other from a one hop mesh perspective. And then clients could also use that. So kind of the same concepts apply in terms of that connection, and how many clients and what they’re doing.

What I would recommend is that if you are doing mesh, and you want to make sure to protect that, what you can do is that you could actually configure it in such a way where clients can’t connect to those mesh links and stuff like that, and not worry about it.

But in a home environment, you just turn that stuff on and just use it. I wouldn’t be worried about it. Hopefully, that helps a little bit. I’ll give you an example. At the San Diego Zoo, we’re doing some one hop mesh links to get to some areas where they just could not– they just didn’t have infrastructure.

Like out of all the access points we installed, we had like three locations. And then out of those three locations, what we did was, interestingly enough, we used a MIST AP for a one hop mesh. And then on the other side of that MIST AP, the relay link, we actually plugged in another AP.

And we use that AP to service the clients. And then we’re just using the other one to backhaul the mesh. Because we wanted to make sure that we were protecting that connection, and not having clients lose it. So there’s a lot of different ways to do it.

– So yeah, do that in your house. Get two AP43s, and use one for the single hop mesh and then use the other one as an AP. I’m sure your family will be fine with two APs just [INAUDIBLE] on the ceiling. Nobody will care.

Zach says, I personally run almost everything in my house. Eero, Eero hive, Ruckus, Hack WRT54s. Zach, you and I are like 100% in alignment. I’ve run all that stuff too. And the two MIST AP41s totally blow them away. Yes, that’s what I want to hear. I like hearing that.

Another question is does mesh consume one band? Man, I don’t know how to answer this one confidently in regards to MIST because I haven’t done any single hop stuff. But remember this, keep this in mind, we have two radios to work with, one in 2.4, one in 5 gigahertz. They’re half duplex.

If we’re doing a single hop mesh with the 5 gigahertz band, that radio has to do both. It has to serve clients and handle backhaul at the same time. Jean, am I correct on that?

– It does, yeah. And it uses 5 gigahertz for that mesh connection. And you can actually allow or not allow the clients to connect to it, but yeah.

– Yep. And you say that you have no ethernet on the top level. Here in my office, I don’t have ethernet up here, either. And man, it’s going to be a huge project to get it here. Oh, Kush says that live demo has an example of a mesh setup. In fact, it is this one right here– mesh relay. So that is a single hop mesh.

Notice, one thing to notice here that in the AP list– where did it go? There it is. Notice that our ethernet port speed is null. There is no ethernet port speed because there’s no ethernet port connected there. So yeah, there’s an example of single hop mesh in action.

And I probably need to bring one of my 41s ones up here to the office or get it close to the office. But hey, we’ve been on a webinar this whole time. And I think it’s worked.

Let’s see. A question from David– if both options are unchecked under support access, organization level, does anyone have access to our organizations? Have we done any security audits that we can share with customers?

– If it’s unchecked, no one has access.

– Not even MIST support staff, right?

– Yeah.

– That can be really handy like my partner account manager. I shipped him a second AP. And it’s nice to go just look in the back of his– just go look at the site, and see what’s going on. I got his permission first, and he checked the box for me, which helps.

Yeah. John, I might do power line. Can’t do [INAUDIBLE] up here because it’s the same cabling problem. But I probably will do power line. Or maybe I’ll just suck it up, get a drywall saw, and get to work. One or the other, we’ll see.

Any other questions before we wrap it up? Daniel says, cut those walls up. Someday. Any APs coming with a 2.5 gigabit or 10 gigabit interface? Zach, that’s a great question.

Right now, we don’t have any plans to do that I know of. And here’s why. With an 80 megahertz wide channel with a 4 by 4 client, which, as far as I know, don’t exist, 4 by 4 clients, that is. You might get close to a gigabit per second, maybe, just maybe.

But in typical enterprise, when we’re running 20 megahertz channels or 40 megahertz channels, go look at that MCS index chart, you’ll see what I mean. It’s very unlikely that we’ll ever clear a gigabit per second per link.

Or do we have 2.5 gigabit per second right now, Jean, or they’re just 1 gigabit per second?

– We do [INAUDIBLE]. On the AP43, it has a multi gig port.

– OK. Got you.

JEAN: Yep. Yep. And you can– hey, if a customer is interested in that, obviously, you can marry that up to an EX switch that supports multi gig.

– Yeah. So it’s there. I think the whole purpose of that, probably, just Joel’s opinion here. I hope this is correct. But I think that’s there just for a check box for RFPs. I don’t know that you’re ever going to actually take advantage of that right now. But what is coming is Wi-Fi 6E.

You probably heard that the FCC has just made a huge move and open up 6 gigahertz for us. We’re probably going to be able to use much wider channels up there. And so yeah, I think when we start seeing 6 gigahertz Wi-Fi, then we might finally break that one gigabit barrier. And we might start to need multi gig ports. So that’s going to be very, very interesting.

Tree minutes left. Any questions before we get out of here?