Ethernet is the most popular networking topology in the world. In this video, you’ll learn about the Ethernet frame, MAC addressing, half- and full-duplex communication, and more.
Here’s a common network configuration for a SOHO, a Small Office/Home Office. You have an internet connection. That internet connection usually has a router. Sometimes, the router and the switch and the access point are inside the same device. But I’ve broken them out to show them differently on this diagram. And you might have some devices that are connected over the wireless access point and others that might be directly connected to the switch using a wired Ethernet cable.
If you’re sending information from one of these devices, you might send it out over the wireless network to the access point. That would then transfer it to the switch, which sends it to a router, which sends it out to the internet. The real challenge for the network administrator is to determine what the route would be for this traffic and determine what type of equipment you would need to be able to provide this connectivity.
On an enterprise network, we add a bit more complexity, and certainly hundreds or even thousands of individual users. They might all be in a large office building with multiple floors. And in this design, I have core switches that are on the first floor. And each one of those switches is connecting the switches on the second, third, and fourth floors.
You can see that those switches are then connected to end users’ web desks on that particular floor. To be able to send traffic from one place to the other, they would then be set up to a particular floor, out to the user, and then sent back again in reversing the direction.
We could even have a campus of these buildings all connected together with these core switches in each individual building. And you can see that the complexity of designing an Ethernet network can increase rapidly, depending on the number of devices, number of locations, and number of infrastructure components that you would add to the network.
To understand more about how traffic gets from one side of the network to the other, it’d be useful to examine what’s inside of an Ethernet frame. I have a representation of an Ethernet frame here at the bottom. And let’s step through each one of the fields that you would find in this frame.
Let’s start with the preamble. The preamble is 7 bytes long. And it’s a series of 56 alternating 1’s and 0’s. This preamble is a notification to a device that everything after the preamble is going to be an Ethernet frame.
Next, we have the beginning of the frame. It has a specific name called the Start Frame Delimiter. It’s 1 byte long. And this designates that the preamble is over and we’re now ready to begin the main part of the Ethernet frame.
The first real Ethernet data we’ll see in this frame is the 6-byte destination MAC address. This is the destination for the data that’s contained within this Ethernet frame.
Right after that destination MAC address is the source MAC address, also a 6-byte frame. And both of those together determine where the traffic is going and where it originated from.
Right after the source MAC address is an EtherType field. This is a small, 2-byte field that describes what comes next inside of this Ethernet frame. We can put all kinds of different data inside of Ethernet. That type field will help the receiving device determine how the data inside of this frame should be handled.
Next is the payload. This is, obviously, not to scale because the payload can be a variable amount of data between 46 and 1,500 bytes if you’re not using jumbo frames. We would put our Layer 3 data inside of that payload. So if you were looking for the IP traffic, it would be contained within that payload field.
And lastly is the FCS, or the Frame Check Sequence. It is a cyclic redundancy check that allows us to make sure that everything that we’ve received in that Ethernet frame has not been corrupted as it was transmitted across the network.
Although you don’t have to rebuild an Ethernet frame as part of your Network+ studies, it’s useful to know that inside of this Ethernet frame are things like the destination MAC address and source MAC address, which are very useful for determining how switching is used on our network. We also have this payload, which is where the IP traffic is. And that’s the bit of data that’s used by routers to be able to send traffic between different IP subnets.
Let’s get more information on these MAC addresses. Inside of this Ethernet frame is a destination MAC address and a source MAC address. That term “MAC” stands for Media Access Control. It is the physical address of your Ethernet adapter. And every device has a unique MAC address. Ideally, there would not be another MAC address in the world that matches the MAC address on your Ethernet card.
The MAC address is 6 bytes long, or 48 bits long. And it’s common to show the MAC address in hexadecimal. For example, this is a very common MAC address, 8c:2d:aa:4b:98:a7. All of that together is an Ethernet MAC address.
It’s not only common to display it in hexadecimal form, but it’s also common to add delimiters make it a bit easier to read. The first 3 bytes of the MAC address are what we call an Organizationally Unique Identifier, or OUI.
You can think of this as the manufacturer ID for this Ethernet adapter. And every manufacturer has an OUI that’s been assigned to them. And in some cases, manufacturers may have multiple OUIs so that when you see the first 3 bytes of a MAC address, you know exactly who made that adapter.
The last 3 bytes of the MAC address are a network interface controller-specific value. You can think of this as the serial number for this Ethernet adapter. Although your Ethernet adapters may have the same 3 bytes because they’re all made by the same manufacturer, you’ll find that all of your MAC addresses will have different last 3 bytes because those are a different serial number for every individual device.
Another concept we work with quite a bit in Ethernet is one of duplex. Duplex describes how we’re able to send data between two devices, either one at a time or at the same time.
Half-duplex is a device that cannot send and receive information at the same time. You’re either sending data or receiving data, and not doing both of those at the same time using half-duplex. If you’re connected to an Ethernet hub, then you’re using a half-duplex device. And there’s no way to perform full-duplex communication with a hub.
Switch interfaces can be configured either as half-duplex or full-duplex. But you will very commonly find most switch connections are automatically configured to always communicate over full-duplex links.
With full-duplex, we’re able to send and receive data at the same time over the same Ethernet interface. This is the most common configuration for switches because it provides us with the highest amount of throughput.
The half-duplex operation of a hub is not one we commonly see these days. But it’s useful for understanding how half-duplex and full-duplex might operate.
Here we have a hub, and we have multiple users connected to that hub. Sam is going to be sending information to the SGC server. But Sam doesn’t know where that server is on the network. All Sam knows is that she’s going to set that traffic to a hub. Since a hub doesn’t have any intelligence or knowledge of where any other device is located on the network, the only way that it can communicate to the SGC server is to send that traffic out every other interface on the switch. Obviously, every device connected to this hub is going to receive a copy of that traffic. But only the SGC server will use that traffic because the destination MAC address is specific to the SGC server.
I mentioned that with half-duplex, you can either send or receive at the same time. But there may be times where two separate devices begin sending traffic simultaneously. And when that occurs, we have a collision.
Let’s take this example, again, of Sam sending information to the SGC server. But at the same time, the SGC server is going to send traffic. And somewhere on that network, those signals will conflict with each other and create a collision.
All of the devices on the network will hear this collision occurring. They will wait a random amount of time and then try to resend that traffic over the network, hopefully preventing a collision from occurring again.
This basic functionality of half-duplex Ethernet is described as CSMA/CD. Let’s break this apart. We’ll start with the CS. This means Carrier Sense.
Whenever you connect to a network, the Ethernet adapter needs to identify that it is connected to a network with other devices. And it senses a carrier that it’s able to use to communicate across that network.
Next is the MA. The MA stands for Multiple Access. And that designates that you can have many different devices communicating on the same Ethernet network.
The CD at the end stands for Collision Detect. And we’ve already talked about how collisions occur on the network and how all of the devices can detect when a collision happens. It’s important that all of the devices are able to detect that collision so that they can all back off, wait a random amount of time, and then try transmitting their data.
CSMA/CD and this idea of detecting collisions is, obviously, only required if you’re on a half-duplex Ethernet connection because with full-duplex Ethernet, there’s no collisions.
From a single workstation’s perspective, CSMA/CD means that it will listen to the network to see if there’s any information currently being transferred. And if there isn’t, it can then send some traffic over that Ethernet network.
Notice that there’s no prioritization. There’s no queue. There’s no check-in process. A device simply listens to see if everything’s clear and then tries to send traffic over that Ethernet connection. This, obviously, means that two devices could communicate at the same time, creating a collision. Those devices will then send a signal called a jam across the network, letting everyone know that a collision has occurred. And it, effectively, clears out the network and lets everyone know to wait a random amount of time before trying to retransmit that data.
Let’s now look at a full-duplex network. And to be able to communicate with full-duplex, we’ll get rid of our hub and we’ll replace that hub with switches.
When Sam sends some traffic over the network, Sam’s MAC address will be the source of that traffic. And it’s going to be sending information to the SGC server. And Sam has already determined what the MAC address is of that server.
Sam is going to send that frame to switch A. Switch A will look through this frame at the destination MAC address, determine that the best way to get to that destination MAC address is to send that frame to switch B.
Switch B will perform the same analysis of this frame. It will look at the destination MAC address. It will identify that the destination in this frame matches the destination of the SGC server that’s currently connected. And it will send that traffic down to the SGC server.
You’ll notice that there was no flooding of frame across the network. We had a very directed frame that went from Sam’s workstation directly to the SGC server. In fact, because this is a full-duplex connection, we can have Sam and the SGC server sending information at the same time over the same switches and the same wires of Ethernet. And they’re able to do that without having any collision because with full-duplex, you can send and receive data simultaneously.