Ethernet is the most common network topology in the world. In this video, you’ll learn the fundamentals of how Ethernet communicates from one device to another and how half-duplex and full-duplex configuration can effect this data transfer.
<< Previous Video: Understanding the OSI Model Next: Network Switching Overview >>
Here’s a picture of what the network commonly looks like in our homes. We have an internet connection, and this connection is usually, connected to a wireless router. And in this case, I’ve broken out the different components of that wireless router to have our wireless access point, the switch interfaces on the back of that wireless router, and the routing function itself.
And of course, we have computers and network devices that are on the inside of our network, all communicating out to the internet and back again.
From our computer, we may send information over the wireless network that then connects to the switch and is ultimately, routed out to the internet. That information goes to a server on the internet side and then, begins the process of coming back inside of our router to our switch interface, onto the wireless network and back to our computer.
It’s not much different in an enterprise network. There may be separate switches, and separate routers, and separate wireless access points, but the functionality is exactly the same functionality that it has inside of your home. You might have a central core of switches on the first floor of your building and it might communicate up to the second floor to your computer. And then, your computer may send that information back down to the first floor again. It’s using exactly the same type of Ethernet network, with exactly the same protocols as it’s using in your home.
And of course, the same process occurs as the network grows even larger. On a large enterprise network, there may be individual buildings with switches inside each one of those, but the process of sending that Ethernet frame across the network and getting a response from that device is exactly the same as it occurs in every other Ethernet network in the world.
On this particular slide, I want to break out for you, what’s inside of that Ethernet frame. This level of detail is not necessarily required for the Network Plus certification, but is useful for understanding the way that our switches and our routers work inside of our network.
If you were to capture an Ethernet frame with a Packet Analyzer and view the components of that frame, it would look a lot like what we have at the bottom of the screen, here. We start with a preamble and then, there’s a start frame delimiter. Then, there’s a destination MAC address– that’s where this particular frame is destined to. And we also have inside of this frame, the source MAC address or the original device that was sending this data across the network.
There’s a type field. The payload itself– whatever we’re sending across to that destination device. And finally, an FCS, or frame check sequence, that examines everything that we sent across the network.
This preamble that we start with is an alternating number of ones and zeros. This is the way that the receiving workstation knows that the rest of the information that’s about to arrive, is part of an Ethernet frame.
Then, we have a start frame delimiter, which is a very specific set of ones and zeros that is 1, 0, 1, 0, 1, 0, 1, 1. And that means that everything after this point is going to be related to this Ethernet communication.
The next two pieces of information are what you might expect to have inside of an Ethernet frame– that’s the destination on where this frame is going and the source address that details who sent this information across the network.
This next piece of information is the Type field. This is also called the ether type, which describes what’s inside of this particular frame. This Ether Type field might describe that there is IP version four traffic in the rest of this frame. Or it might say there is IP version six traffic in the rest of this frame.
And then, finally, is the payload. There might be IP information inside of this frame. It might also have TCP data. And it may be sending browsing information. But all of that is contained within the payload of the Ethernet frame.
And lastly, is the frame check sequence. This is something that is part of every Ethernet frame that goes across the network. The frame check sequence is a CRC checksum that looks at all of the data that was sent in the Ethernet frame and it’s able to come up with a single checksum. When your device receives this frame, it calculates the frame check sequence in exactly the same way that the original device did. And then, it compares that time frame check sequence to what’s inside of the frame itself.
If those two matched, then the frame was received correctly. If this frame check sequence doesn’t match, then we know that something was corrupted with this data somewhere along the path. Your Ethernet card will drop this frame, and it will increment a counter inside of the Ethernet card for CRC errors.
So if you examine the CRC counter on your Ethernet card, you’ll be able to tell how many frames were dropped because the frame check sequence did not match.
We have already talked a bit in this course about the MAC address that’s associated with an Ethernet adapter. This is called the Media Access Control address, and many people refer to this as the physical address of that Ethernet adapter. The MAC address that’s associated with your Ethernet card is an unique address. There’s no other device in the world that has exactly that same MAC address. That means on your local network, every device can be communicated with using this unique Media Access Control address.
The MAC address is 48 bits long or six bytes long, if we do the math. And we usually, display a MAC address in a hexadecimal. So this MAC address is 8c, 2d, aa, 4b, 98, a7, and we’ll usually put some type of delimiter between those hexadecimal values so that we can read it just a little bit easier. It might be colons or we might use dashes.
The MAC address is separated into two pieces. The first half of the MAC address is the OUI or the organizationally unique identifier. This is a value that’s associated with the manufacturer of this Ethernet adapter and that means we should be able to look at those first three bytes of a MAC address and determine who’s the manufacturer of that particular adapter card.
The last three bytes are the ones that are associated with that individual card. You can think of this as the serial number of this particular Ethernet card. And with those six bytes together, we know that there’s no other device in the world that has exactly that same MAC address.
If you’re connecting your Ethernet device to a network, you may see some configuration parameters that are specifying which duplex you’d like to use and usually, you have the choice between half duplex and full duplex. With half duplex, your device cannot send and receive data simultaneously. It has to either, receive data or it has to either send data, but you can’t do both of those at the same time.
If you ever find an old Ethernet hub lying around, those are half duplex devices. And you would not be able to communicate using full duplex if you were using an old hub.
These days, we use switches. Switches, although can be configured as half duplex devices, it’s much more common to use a switch that’s configured as full duplex. That’s because the only thing connected to a switch is your one device. That means that we can send data to the switch and receive data from the switch simultaneously. This is full duplex communication, but you have to make sure both the switch and the end station are both configured for full duplex.
If you did manage to go back in time and find a network that had a hub and had devices connected to that hub, they would all be communicating via half duplex. This means that we have all of these devices and, let’s say– in this case– that Sam is communicating to the SGC server. Sam would send that Ethernet traffic to the hub. The hub then sends all of that traffic to every other device on the network.
A hub doesn’t have a way to intelligently forward traffic so when it’s communicating across the network, it’s sending all traffic to all devices all the time.
On this half duplex Ethernet, there could be situations where two devices communicate simultaneously. In that particular case, you have what’s known as a collision. Let’s say, that Sam is sending information to the SGC server and the SGC server is sending information out at exactly the same time. Those two frames create a collision, and everybody on the network will hear that collision when it happens. Everyone waits a random amount of time and then, tries to communicate, again, and hopes that a collision doesn’t occur a second time.
We call this type of half duplex communication CSMA/CD. The CS in CSMA, means carrier sense. That means that your Ethernet adapter is going to listen to see if there’s a carrier that’s available that it can use to send a frame of data out to the network. The MA in CSMA/CD, stands for multiple access. And it means that there’s more than one device that is on this particular network.
The CD in CSMA/CD, stands for collision detect. That means if two devices do communicate simultaneously and that information does collide, those devices will recognize that a collision has occurred. They’ll identify that that data was not able to get through. And then, they will perform a random backoff function. And then, try to communicate that traffic across the network again.
You commonly CSMA/CD referenced when you’re referring to a half duplex internet communication. On today’s modern switch networks, almost all devices will be configured as a full duplex device. And they will not be using CSMA/CD.
On a shared half duplex network, you have a CSMA/CD operation that starts with a device listening to make sure that it doesn’t hear anyone else communicating on the network. Obviously, you don’t want to transmit information if there’s already traffic going across the network. If the network is clear, device will try to send data across the network.
There’s no que. There’s no prioritization. Devices simply try to send information, if the signal happens to be clear.
If two devices do happen to send data simultaneously, there’s going to be a collision. And both of those devices will hear the collision and they’ll send what’s called a jam signal across the network, that effectively clears out the network and lets everyone know that a collision has occurred. All of the devices then wait a random amount of time and then, try to send that data again.
On a switch network, the switch makes the determination intelligently on where the traffic happens to go. Unlike a hub that sends traffic to all of the interfaces, a switch will find exactly where the traffic needs to go and send it just to that particular device.
Let’s take the example of Sam on this network, sending information to the SGC server. Sam is going to send an Ethernet frame that has Sam’s MAC address as the source MAC address and the SGC server as the destination MAC address. Sam’s connected to switch A so Sam sends that time frame out to the switch. Switch A is then going to examine where the destination is for that SGC server and know that the only way you can get there, is sending it out on this interface to switch B.
Switch B then performs exactly that same lookup function. It will look at the destination MAC address, find where that device is located on the switch, and send that Ethernet frame down to that device.
On a full duplex Ethernet network, that function can occur in both directions simultaneously. So Sam can be sending data to the SGC server, while at the exactly same time, the SGC server can be sending information to Sam. Since this is a full duplex network, there’s no way that the frames could possibly collide with each other.
On a wireless network, the communication process works a little bit differently. On wireless networks, you’ll often use CSMA.CA. The CA stands for collision avoidance. On a wireless network, as a device is sending information, it’s effectively overloading its receiver so it’s not able to hear anything else while it’s transmitting data.
Because of that, wireless networks may use the CSMA/CA function where it will use ready to send and clear to send, to be able to send data. A device may communicate to an access point and say that I am now ready to send. The access point will give that particular device permission to send. And only that device can send during that time frame. That device will then send its data and then, the next device will have to ask for permission from the access point.
This solves a problem on wireless networks called the hidden node problem. You might have an access point. There might be device A on one side of the access point and device B on the other side of the access point. Station A can hear the access point. Station B can hear the access point. But station A cannot hear station B because they’re too far from each other.
With this ready to send, clear to send functionality, both device A and device B can communicate on the same wireless network even though both of those devices cannot hear the wireless communication from each other.