CWNP CWNA – Wi -Fi Access

1. Wi -Fi Access

Now in this module we’re going to talk about WiFi access and we’re going to make a couple of comparisons about how wireless is different than a wired. So we’re going to compare this idea of CSMA carrier sens multiple access with the WiFi using collision avoidance and with the wired using the collision detection. Then we’re going to get into some more acronyms which I’m sure all of you are waiting for. We’ll talk about the distribution distributed coordination function, the point coordinated function, the hybrid coordination function that’s getting harder to say block acknowledgments, the WiFi multimedia and then talk about this concept of airtime fairness.

2. CSMA/CA vs. CSMA/CD

So when we make the comparison between collision avoidance and collision detection, collision avoidance is what we use. Actually. We use it for both wired and wireless. At least we used to. Today’s world, we don’t worry about collisions in the wired network because we’re not doing anything. Very often that’s involving half duplex. But you see, that is the problem with wireless though, is that with half duplex that means you can either transmit or you can receive.

But you can’t do both at the same time. And certainly if I have two computers on the same channel on the same frequency and they both want to speak at the same time and it’s radio frequency, they’re going to collide. And whatever makes it the access point is not going to make sense. So we have to find a way to avoid those collisions.

Now, a lot of this is done with what we call the media access control, which is a generic term that we use for the general concept of having access wireless or wired into the network. Now, like I said, the two most used are the CSMA, which is a carrier sense meaning listening to hear. If it’s clear to send traffic multiple access, we’re not given terms to go. We just go when we’re supposed to. And then it’s about whether or not we’re looking at collision detections or collision avoidance.

3. CSMA/CD (eNotes)

In the old days. I say old days, right, the nineties and a little bit earlier we were so happy to buy this little product called a hub. And a hub was like a power outlet. I hope that makes sense. Something you plug into the wall and it’s a cord and it has like six or more different places for you to plug other stuff, stuff in. So they’re all sharing the same electrical current that’s coming through that. And that’s what a hub did is it allowed us to plug multiple computers in and share the connection from what we often called an uplink connection to some other part of the local area network.

But internally you were sharing the bandwidth. So if I had a computer on one port and I put another computer on another port they would all be able to get to the rest of your local area network. But they had to take turns. They couldn’t both do it at the same time. Back again into the days when we had what were called party lines. And I know I’m aging myself when I talk about these things but there was a day, and I think I mentioned this before, that we had phones at our house that were plugged into the wall and we could have more than one phone but we all had to share the same one phone line. And that’s kind of what this is doing here today. It’s different. Nobody has a phone plugged into the wall and they’re all using their own cell phones. That was also a day in the United States when I only had to call or use or dial four numbers to be able to talk to my grandmother.

I didn’t have to worry about the area code or the city prefix or anything else. I mean, times are different back then, maybe simpler. But besides my going back in time, the idea here is that they would have collisions. If one computer decided to talk at the same time another computer decided to talk, both of their electrical signals would collide and whatever came out of it to go to the local area network would just be a mess.

And so we had to have a way of doing collision detection. So the first part was the carrier sense. The carrier sense is where we’re basically having both computers listening to the traffic. And if they don’t hear any traffic then they are going to assume that they’re going to have what’s called a clear to send signal. Now they both can listen at the same time or they are listening at the same time and they both might choose for whatever reason at the same time to begin a transmission which is where the Ma, the multiple access came in.

But if they did somehow interfere with each other’s communications then the CD, the collision detection would come into play. So what would happen then is if I had computer one send something at the same time. Computer two did. Because they listened, it was clear to send it’s multiple access, no turns needed, and then the resultant collision. They would both detect the mess. Each one of them would pick a random amount of time. If you can’t tell us a clock, they would pick a random amount of time called a back off timer, and basically, who’s ever clock expired and went to zero first would be allowed to speak first. I mean, they would be able to do their transmission. And when the other one’s time got to zero, they would still not be able to transmit because the first computer is still doing its transmission, but they would no longer collide.

Now, the bad thing is they could both have their back off timers, but if computer three said, hey, you guys aren’t talking, and that’s because they’re waiting, then it might set a transmission off and just make a mess. So the more computers we added to this, the more we were sharing our bandwidth. Now, hubs acted in also in half duplex, meaning, again, you could either transmit or you could receive, but you could not do both at the same time. Now, today we have a device that almost looks the same as a hub, at least if you were to look at it from the front on a rack. Other than that, this device usually has many more ports, and today we call them switches.

And so on a switch, it gave everybody a full duplex connection. And by full duplex, it meant that you could speak and speak and receive at the same time. So think about it. Let’s say on the Hub that you’re listening. You don’t hear any communications, so you begin to send traffic, but at the same time, somebody’s sending you traffic. Well, even on that wire, you could have a collision and lose your data because of half duplex. So we don’t have any collisions on the wire. In fact, what they tell us is that every port on a switch and now switch uses different kinds of arrows that you have your own collision domain.

So now you can send your traffic, you can listen for traffic. And if you both send at the same time, the switch stores inside of itself in a memory area you might call a buffer if it needs to store it at all to be able to afford that traffic without causing any collisions.

4. CSMA/CA Part1

With wireless, things are a little bit different. And don’t get me wrong, on the wired networks there used to be this thing called collision avoidance, but I’m going to describe it with the wireless rather than on the wired network. So Csmaca is a process we use to make sure that only one 8211 radio is transmitting at a time. Now, this is a part of the IEEE 8211 2012 standard definition that defines this function called a distributed coordination function. Distributed meaning many different clients trying to make communications at the access point. So we’re coordinating how that works.

It’s a medium access method that lets us have a lot of different types of checks and balances to try to minimize the chance of collisions. Now, this process does not really tell you if they’re was a collision like collision detection did. In other words, there is no way to find out if there’s a collision other than the fact that you’ll find out that most every frame or bit of data that you send, you’re going to receive an Acknowledgment to let you know that it was received. And if you don’t get the Acknowledgment, the assumption is your traffic had a collision somewhere.

5. CSMA/CA Part2

Now, a part of the DCF is what we call the fundamental access method, which is a part of the 800 and 211 communications. And it is the mandatory access method for anybody that’s following the standards that basically says that if you send traffic, you have to have that Acknowledgment. If you don’t have the Acknowledgment, then you are going to assume that you had a collision with your traffic.

6. CSMA/CA (eNotes)

So let’s take a look at this a little bit more in detail. Let me put an access point over here and I hope you don’t mind if I don’t draw what looks like radio waves coming out there. And I have a lot of computers that are talking to this access point. But we have to remember that when every computer talks, everybody else in range is going to hear that conversation.

And the access point might also want of talk to send the clients information. But because it is half duplex, everybody has to take a turn. So they still do carrier sense where they’re listening. They’re listening to see if it’s clear to send traffic. And if it is clear to send traffic, then if everybody’s listening and they all know it’s clear, they’re going to use multiple access, which means we’re not giving everybody a turn. We’re not like token ring where computer A gets to go first, then computer B, then computer C whether you have anything to say or not. We just manually give or through the software give everybody a turn. It’s multiple access. Anybody can start. The thing that’s different is the collision avoidance.

So let’s say that all these computers are listening, they realize it’s clear to sand. One of them is going to send a packet. And that’s what I’m going to call the collision avoidance packet. It’s only a packet to get permission from the access point to be able to speak. Now because I sent that packet, everybody in the range of my radio is going to hear that.

So they’re going to be quiet because they realized, hey, we were a little bit too slow to send that packet so we have to wait our turn. We’re going to get into a little bit more detail about that process as we go on. But the idea is to get basically the access point to say okay, I hear you, I know you want to send it’s your turn. I’ll acknowledge that it’s your turn and then you can start sending all your data and everybody is going to wait for their turn.

Now it actually is up to the access point to pretty much at some point say hey, you know what, you’ve had enough time. It’s somebody else’s turn to talk. And by the way, this is happening in microseconds. So even though it sounds like a lot of work, sounds like a lot of labor to be able to transmit to our perspective as people, it’s happening almost instantaneously.

7. Distributed Coordination Function Part1

So the DCF that distributed coordination function is a fundamental access method of 800 and 211 communications. We again said it’s mandatory. Now, there are some other types of optional access methods that we could use. Some that you’ll see a little bit more of, like the PCF or the point coordination function or even another one, is called the HCF, called the hybrid coordination function. Those are things that have been added. But the DCF, as I said, is absolutely mandatory to have.

8. Distributed Coordination Function Part2

So as we continue to talk about the DCF, some of the things that we have to look at is the components that make up the DCF is really a part of that whole Carrier Sense multiple access collision avoidance process, the part that we said was absolutely mandatory. Now there are a couple of other components of that to make this all work. And we’ll get a chance to talk about more of these in detail, but we’re going going to at least give you some kind of an idea.

So one of the things I talked about was the idea of Carrier Sense. Now when we get to Carrier Sense, we’ll talk about the physical and the virtual part of how the different clients are listening to know if it’s clear to send something. And you have to remember, again, in the world of radio, whenever your client transmits, all the other clients are going to hear that. And so one of the other things you’re going to hear is what’s called a duration or ID field, which is really when the computer that has its turn to send is going to send its packet and it’s going to say, this is how many milliseconds it’s going to take me to send you the data that I want to send you.

So all of the rest of the clients are like, okay, let’s write down that number. So we know that we have to wait at least that long before it’s our term. If we don’t get an acknowledgment, some of them might have a random back off timer like we do with collision detection, where they just automatically wait a random amount of time before they listen and try again. And then there’s also this thing called the interframe space. Now you’re going to hear about the interframe many different times with things like SYFS and diffs. So the if interframe is going to have different types of timers that indicate something different, and we’re going to talk about those as well.

But those are components that everybody hears. And by hearing it, then we have, I guess, maybe a little bit better coordination between all of the clients that are connected to that access point. So we don’t really have a lot of collisions. So basically, think of these four components I just mentioned as checks and balances to work together to try to keep from retransmitting the same information over and over again and causing everybody else’s problems. And the reason we need them is radio frequency is a half duplex medium.

9. Interframe Space (IFS)

So the interframe space, the ifs the thing that I just talked about is basically just a period of time that we use to indicate what time are going to be occurring between the transmission of different wireless frames. There are six different types of these interframe spaces. The majority of them, over 95% of them are going to be the sifts and diffs that you’ll talk about later.

But there are a few others in there for some specific little things. Like one is the reduced interframe space. It gets, by the way, the highest priority. We might use that for quality of service. If somebody’s making a phone call with an IP phone over a wireless network, that traffic is really bad with latency, it just doesn’t work. So it gets the highest priority. The next one, the SIFs, the short interframe space, again, second highest in priority ones you will very rarely ever have to deal with, like the PCF or the PIFFS.

The PCF interframe space, middle priority. Then the other one, the diffs that you will see a lot is called the DCF interframe space. And we already talked about what DCF, excuse me, is that whole issue that we just went through with the collision avoidance lowest priority. There’s also an arbitration interframe space that’s also used, by the way, with the QoS stations.

And the arbitration is about, again, getting more of a priority for your traffic because it is affected by latency. Latency meaning it takes a long time from the sender to the receiver. And when I say receiver, I don’t mean the access point, I mean from host to host to be able to get that communication. So we want to give it the appropriate time. And then there’s the extended interface or interframe space or the EIFS, also used to basically help us understand when we get corrupted frames.

10. SIFS/DIFS Part1

So when we look at the sys and diffs, one of the examples that we might see are things like the AC frames, which is the Acknowledgment. We’ll talk a little bit later about the block Acknowledgment frames. Obviously, we’re sending data there, and we also need to know when it’s clear to send. And these frames may follow what we call that SYFS. And so when we look at this kind of a diagram here, here we see that maybe we started by sending some data.

And while the nose transmitting, all the other wireless devices hopefully hear that, and they don’t wait or sorry, they don’t transmit because they know they have to wait. Now, there’s going to be a short interframe time that occurs before the Acknowledgment gets sent back to the person that sent the data. And that’s just a normal part.

So if we had that duration, we said, hey, this is how long it’s going to take me to send you that data. Everybody else listening says, okay, I know how long they need to send the frame. I also know that I’m going to add the Sifts time, which means that by adding it, it means everybody knows just because they’re done with that packet or that frame not to send, they got to wait for that time to send the Acknowledgment.

And then after all of that’s done, the station must wait for at least two other time periods that we call the diffs, which occurs after the acknowledgement when they decide, okay, I don’t hear them talking. I’ve waited the amount of time you told me to wait. So now I’m going to send my data because it sent out, seems to, or appears to be my turn to be able to send. So it’s just about, again, a coordination.