Time
31 hours 29 minutes
Difficulty
Beginner
CEU/CPE
30

Video Description

LAN Technologies: Broadcast, Collision, CSMACD and CSMACA This lesson discusses the technologies of Local Area Networks (LANs). The technologies of Local Area Networks include:

  • Broadcast: sends to all, only on one side of router
  • Collision: two devices transmit simultaneously, data on same line collides and corrupts, there needs to be ways to avoid collisions.
  • CSMA/CD: carrier sense, multiple access, collision detection.
  • CSMA/CA: carrier sense, multiple access, collision avoidance

Video Transcription

00:04
So what? We're talking about the technologies on our local area networks. The type of traffic that's going on on our local area networks is quite different than the traffic that we have going on on wide area networks on our local area networks. We have multiple devices that can talk, try to talk at the same time that are trying to talk to multiple different points.
00:21
Where I was on our wide area networks, Things are a bit more.
00:24
They're a bit more organized. They're a bit more point the point if you're transmitting over a wide area network, typically you're directing your packet directly to a particular in point on our local area. Networks weaken do things called broadcast
00:40
now, whereas on our wide area, networks were typically sending a packet from us to a server
00:46
or from us to another computer, a packet directly from point A to point B. On our when we're connected to a local area network, we may not know what where to go. We may not have an I P address yet, so we have to send out a broadcast packet. We've talked about broadcast packets before as being packets that we send out that are to everyone that we can connect to.
01:04
These are very common when we're doing things such as art address, resolution protocol or we're doing DIY HCP, where we're connecting to a network and we're doing a D h e P Discover or a D A C P request and saying, Hey, can I use this I p address or Hey, is there anyone out here that does the HCP? We're sending a broadcast message out to everybody
01:26
now broadcast messages on Lee propagate on our side of the router. So
01:33
if we have a network where we have three computers connected to a switch connected to a router, then if one of these computers sins out a broadcast message, that broadcast message will go to everyone.
01:49
But the router will not foretell Ford It alone. Hubs and switches and bridges, hubs and switches and bridges will push along broadcast messages, but routers will not push along broadcast messages. Just think of the havoc that could, because that they would if you could push out a broadcast message and routers would pass that along,
02:08
you would have people on the other side of the world getting your broadcast messages because it would just propagate through every network everywhere and just keep going. Eso broadcast messages air not pushed through out past a router.
02:21
Um,
02:23
so this is where we end up with what we call a broadcast domain are broadcast domain is everywhere in, connected to a single point where a broadcast message will reach. So if we are, If we were to define our broadcast domain in this network, our broadcast domain would pretty much be everything,
02:43
including the port on our side of the router.
02:46
So when we're looking at a network topology, we're looking at a network map, and then we're defining our different broadcast domains were defining all of the computers server switches that are connected to each other and are on the same side of a router on the same sub net on the same violent. So that's going to be our broadcast domain.
03:06
And we also have something called a broadcast storm. Now, broadcast storm occurs when we have miss configuration of devices, and a broadcast storm is essentially an endless loop of broadcast messages passed around over and over and over that can effectively shut down our networks.
03:22
Now
03:24
what we're gonna do is we're going to throw in in our little diagram.
03:29
Two other switches
03:31
and all three of our switches are formed together. Two are connected together to form a loop.
03:38
And then we'll say that each of these switches air connected to additional computers and one of them is connected to a router.
03:44
Well, if we send a broadcast message to one of our switches that switches going to propagate the broadcast message to the next switch which will push it to the next switch which will put you back to the first switch, which will put which will push it to the second switch. And so are switches will essentially keep pushing that broadcast, pack it around to each other
04:03
and keep pushing them out to their computers
04:05
over and over and over and over, and it will end.
04:08
And that will cause essentially a shutdown on our network. Because we have so many packets traveling over our network that is causing collisions,
04:15
so broadcast storm will broadcast storm will essentially shut down our network due to collisions.
04:21
So what are collisions?
04:25
Well, a collision is what occurs. We have two devices that it tried to transmit over a over cable simultaneously. Our network cables are designed to only be able to send and receive from one packet at a time. If we have two devices that are trying to transmit simultaneously,
04:44
the electrical signals will actually hit each other on that cable.
04:46
And we'll we'll form what's caused a collision which will corrupt the packets. And neither of those packets can get to their destinations.
04:56
So if we have,
04:57
say, computer, eh
05:00
and computer be
05:02
and they're both connected to a switch
05:04
are actually they're both connected to Yeah, they're both connected to a switch.
05:10
And let's say that they're both trying to send data down to computer. See?
05:15
Well, our switch will help mitigate a collision. Are switch will help mitigate a collision.
05:21
But if
05:23
say, that switch is trying to transmit to them
05:27
over the link, the switch is trying to talk to them.
05:30
At the same time. They're trying to talk to the switch in, the data collides.
05:34
Then it'll cause a problem. Now the window for this collision isn't
05:40
two or three seconds are windows, for our collisions are very small. They're there in the milliseconds range because our data transmits over our cables very, very fast. But if in the case data does collide, it will become corrupted, and it can't send our can't keep going, that's what That's why our broadcast storms are such a big deal.
06:00
Because we have so much data flooding our network
06:02
that the collisions just keep happening no matter what, because they're all they just keep traveling around our network.
06:08
Um,
06:10
now,
06:11
when we have, we have data, Kalai data collisions that corrupts our packets. Those packets don't just disappear. Are devices that sent the data can recognize when a collision occurs and, well, actually
06:24
perform either CSM a CD or C S. M a c A. Which we'll talk about in a little bit that allows us to rescind that data packet if we detect a collision.
06:34
So it's important that we have a way that we can mitigate these collisions.
06:40
Now we have written here that we need to talk about the difference between our collision domain and our broadcast. The main. This is a topic that we've covered in a previous module. So we will talk about it a little bit briefly, but are a collision domain in our broadcast domain is essentially the areas of our network. We've mentioned that our broadcast domain is the areas in our network
07:00
where a broadcast message will propagate
07:02
our collision. Domains are links where collisions can occur from a single computer
07:10
example.
07:13
So we have our network here
07:14
with our three switches, and we're going to say that we implement it implements it spanning tree protocol. So these three switches are no longer creating broadcast storms,
07:26
but they are all in the same broadcast domain.
07:30
Well, while they are in the same broadcast domain, we now have several sub
07:34
collision domains. Because all of these devices are switches. Remember, Switches will on Lee sin packets to their intended recipient. They won't push packets out to everybody like a hub.
07:46
So the only time will experience a collision is if our switch is trying to talk to our computer at the same time. So if we're using switches, then our collision domains are smaller because our collision domains are poor. Areas where collisions can occur are only going to be us connected to are connected to our in point.
08:07
Now, if this device was not a switch, if this was device was a hub remember, hubs will receive a message from one computer and just send it to everyone that they're connected. Thio.
08:18
So if this computer was to send a packet to our hug, it would send it to everybody else.
08:24
So that would now make our collision domain much larger. Because we send a single packet, that pack it's going to send to everybody else.
08:33
Our potential for collisions that this one computer can cause is much bigger if we use switches. The area that are compute our single computer can cause collisions in is much smaller.
08:45
So that's our difference between our broadcast domains and our collision domains are broadcast domains are all of our devices that a broadcast message can reach and our collision domains are all of the devices that are computer can effectively caused data collisions with,
09:01
So we need ways to avoid collisions.
09:05
We need a way thio. We need some sort of standard that allows us to prevent or to recent packets of collisions occur. Now our Ethernet standard are IEEE 802.3 actually includes specifications for avoiding in mitigating collisions.
09:24
This is called R C s M. A slash CD and R C s m a slash, See A So what does that stand for?
09:30
Well, R C s m a slash CD stands for carrier sense multiple access. Collision detection must break this down.
09:39
Our carrier sense
09:41
refers to our nodes. So our computers are switches are devices
09:48
sensing and detecting collisions on our network There there are different devices are sensing if there's collisions,
09:56
multiple access means that we have multiple devices that are trying to talk on our network simultaneously. So why does this matter? Well, if we think about some of our net, some of our network apologies such a czar, mesh or start Apologies versus other apologies, like our rink apologies. Some of our networks don't provide
10:16
simultaneous access to the network at all times for all devices.
10:20
Our ring that our ring network, for example, in our ring topology a device can only talk if it has a clear token. So this is not a multiple access network. Not all our devices can't talk whenever they want to. They have to wait for that token. Our mesh network is a multiple access network.
10:39
Our bus network, our bust apology, or we have all the devices connecting toe. One line
10:46
is a multiple access network. So our networks, where any of our devices can talk whenever they feel like it, they don't have to wait for a token is going to be multiple access. So this CSM a slash CD wouldn't apply in a ring network because each of our devices doesn't wait, are doesn't
11:05
transmit whenever at once. It has the weight. We don't have multiple active devices that can access the network at the same time. And then the last part of this is going to be our collision detection
11:15
Now with R C s M a slash CD What's going to happen when our devices what's gonna happen when our devices detected collision? We have our two devices on our network. Lets say we have a We have a bust apology. So we have our single cable
11:28
with our Terminator.
11:31
We have computer one
11:35
computer too, and our server
11:39
So
11:41
Computer one wants to talk to our server
11:45
at the same time.
11:48
That computer to is trying to send a message to our server.
11:54
So they're going to send data down,
11:56
and that data is goingto form a collision. So we have we have a date a collision
12:03
because we have carrier sense
12:07
both computer one and computer to detect that their data formed a collision on the network. Their data didn't get where it needed to be. There was a collision.
12:16
So what are devices air going to do is they're going to set a random interval of time
12:20
when they detect that collision. Collision detection. They're going to set a random interval of time toe wait before they try to rescind that packet.
12:31
Now, because it's such a because the designation that we use for how long that random interval of time is, which is in milliseconds. So it's
12:39
to a point where, unless we have a lot of collisions going on in our network, we don't really even notice it that much.
12:48
Computer 1 may query. It's processor and say, Okay, well, I need to wait a random period of time. Give me a random integer and the random integer that it gets toe wait is 300 milliseconds.
13:01
So Computer one is going to wait 300 milliseconds to retransmit computer to queries its processor gets a random integer and computer to says, Okay, I'm goingto wait 168 milliseconds before I try toe
13:16
retransmit so one computer to wait 168 milliseconds
13:20
that 121 that 168 milliseconds. Platt passes in the blink of an eye
13:24
computer to re transmit successfully because there's no collision computer one waits a couple more milliseconds to blinks of an eye, and it re trans transmit successfully.
13:37
So
13:37
carrier, since our devices listen. And since when there's a collision
13:43
multiple access, we have a bus network so all of our devices can talk whenever they want to. They could talk at the same time, which has the potential to create a collision.
13:52
Collision detection. As soon as our devices detect that there's a collision, they're goingto wait a random period of time and then try to retransmit
14:01
so
14:03
with R. C s M. A slash CD. This is really a reactive measure. We're just transmitting whenever we want. And then as soon as we're transmitting, we're seeing if that data collides, Well,
14:15
if we're sent, if we're trying to send a lot of data or for sending really large packet frames, that's a lot of data that we're gonna have to retransmit which in terms of which, in terms of our time, where we reference time in seconds and minutes may not seem very long. Maybe a couple extra 100 milliseconds. But for computers,
14:35
the time that it's gonna take for it to re transmit a large packet
14:37
is forever. Long computers work in computers, work much faster on computers, think at a much faster rate and evaluate time differently than we do per se.
14:50
So it's going to see that as a long time that it has to wait to resend a large packet with CSM a slash CD.
14:58
So another option that we have is CSM a slash c a.
15:03
Now c s m a slash see a stands for carrier since multiple access collision avoidance. So we're actually taking proactive measures to prevent collisions from occurring. Now see a is very, very similar to CD in the fact that what's going to happen is our devices are still going to send out a packet
15:22
and see if they can hear a collision and then wait of random period of time if there is a collision.
15:28
But with CSM a slash, see a with collision avoidance. When we're sending that packet to see if the coast is clear. We're not sending a full data packet. What we're sending is were actually sending an alert out to the network that we are about to transmit.
15:46
So
15:46
we clear up our network.
15:48
We have C S m a slash. We have, ah, see a collision avoidance implemented
15:56
and computer A realizes that it needs to talk to the server.
16:02
So it's going to send out an all clear alert message
16:06
all the way out of all the way down our network. That's a very small packet size. So it takes it just a blink of an eye. Very, very short amount of time, even in computer terms, to send that alert packet to say, Hey, I'm about to send a message.
16:19
So whenever anyone transmit for the next couple milliseconds, I'm about to send a message.
16:25
As soon as that our computer realizes that there was no collision.
16:27
It's in this packet,
16:30
so alert, and then it sends the data.
16:33
Now let's say that
16:36
computer to sins and alert at the same time Computer One does, or
16:41
half 1/100 of a millisecond after so that a collision occurs.
16:47
So computer one and computer to don't have a date. A collision, but they have a alert message collision.
16:55
So,
16:56
computer once in that on alert computer to send out an alert. Hey, I'm about to send a message.
17:00
That alert had a collision.
17:03
So
17:03
Computer One and Computer two are both gonna have their back off time.
17:07
They're both gonna wait a random period of time to re transmit that alert message.
17:14
But that's still better than our CD. Because now, even though they're still waiting that random period of time, it takes them less time to re transmit that small alert message than it does to try to re transmit an entire data packet.
17:29
So
17:30
that alert message had a collision. They both wait. They're random back off time computer to is gonna rescind an alert message and say, Hey, I'm going to try again. Here's an alert message. There's no collision this time. And so once it had the old once the had the all clear from the alert message, it's gonna send out its data
17:48
and then a couple milliseconds. After that, our computer one hears that alert message so it holds off, and then after the coast is clear, it will send out its alert message and then it will send its data.
18:02
So
18:03
that's our difference between RCs and May and R. C M S r. C s M A C A r C s m A C D r Collision avoidance and our collision detection
18:11
in both cases are devices are listening for collisions, and there they have a back off time and a recent a re transmit if there is a collision. But remember, in C S m a c a avoidance also, think of that, eh? Is being our alert message.
18:30
Our computers are going to send out a small alert packet
18:33
before they actually transmit their data because it's easier and quicker for them to re transmit that alert packet. Then it would be to retransmit their entire data packet. So with R C M. C S M. A collision avoidance, we may notice that even if there's just as many collisions on our network, it's
18:48
that takes our computers less time to resolve those collisions
18:52
because the packets are smaller

Up Next

CompTIA Network+

This CompTIA Network+ certification training provides you with the knowledge to begin a career in network administration. This online course teaches the skills needed to create, configure, manage, and troubleshoot wireless and wired networks.

Instructed By

Instructor Profile Image
Anthony Harris
Senior Systems Engineer at ZenPoint Solutions
Instructor