There are different configuration options to consider when connecting to an Ethernet network. In this video, you’ll learn about interface settings, link aggregation, MTU sizes, and jumbo frames.
When you’re configuring a router interface, a switch interface, or an interface that’s in your local computer, there are a number of options available. In this video, we’ll look at some of those interface configurations.
We can start with the basic connectivity that we have to the network. When we plug-in an ethernet cable, there are a number of ethernet settings we need to consider, and two of the most important are the ethernet speed and the duplex. The speed is obviously going to be how fast this connection is going to run. This might be a 10 megabit connection, 100 megabits, 1000 megabits, which would also be 1 gig connection, 10 gig or even faster.
We have to be sure that this configuration is the same on both sides. So if we have a computer that’s connecting to a switch, the speed on both of them has to be configured the same. In many cases, we would set this to automatically configure itself so you can simply plug in any device and the switch will automatically adjust to the configured speed.
The other option we might have available is either half duplex or full duplex. In most cases, this is also set to automatic. But if you do set these manually on both of these devices, you have to make sure that both of them are set to be the same.
If you have the speeds mismatched on both sides, the connection will not work at all. You’ll never get a link light. But if you mix the duplex configuration where one side is set to half duplex and the other side is set to full duplex, you’ll find the connection does operate, but you’ll notice that it is not quite as efficient as it should be. And, in fact, you’ll have very poor performance if you start putting a lot of traffic over that link.
This is why one of the initial configurations will look for when we’re troubleshooting ethernet is the speed and the duplex and make sure that they’re the same on both sides of the connection. If those match, then we’ll shift our focus to the internet protocol or the IP configuration for that interface. This might be an ethernet connection, it might be a VLAN config, we might be setting up management interfaces, and all of these will need an IP address, a subnet mask, a default gateway, and perhaps other IP configurations as well.
These will usually be values that have been assigned by a network administrator. So we’ll want to configure and make sure that the IP address, subnet mask, default gateway, DNS settings, and everything else matches the configuration that we’ve received from our admin. If this configuration doesn’t match what we’ve been given, let’s say we’ve used a different default gateway IP address or we put in the incorrect subnet mask, then we may find that we’re not able to connect to other IP devices on this network.
Fortunately, this is a relatively easy configuration to validate. We can simply look at what we’ve been given and look at what we’ve configured in the system and make sure that both of those match.
Another useful configuration, especially between switches, would be one for link aggregation. This is sometimes referred to as port bonding, or you might see it abbreviated as LAG for link aggregation. This means that we are connecting multiple interfaces together on these two devices, which normally we would not want to do because we would be creating a loop. But if you’ve configured all of those interfaces as link aggregation, the switch will interpret them as one very large connection rather than four individual connections.
So if all four of these were gigabit ethernet connections between this switch, we would configure those as LAG and you now have four gigabits of throughput to be able to communicate between those two devices.
Some switches will support an automatic configuration of this link aggregation using a protocol called LACP. This is the Link Aggregation Control Protocol. This allows you to simply configure the interface as LACP, plug-in the interfaces and all of the underlying link aggregation configs will automatically configure themselves between those two devices. If you were to capture traffic between these two devices, you would see LACP traffic being transferred from one switch to the other.
To be able to have the most efficient communication from one device to another, we need to be able to send packets across that will not be fragmented along the way. Since we don’t always know what networks are located between the two devices, there may be times when our data is fragmented and therefore is less efficient to transfer across the network. Fragmenting a packet tends to slow down the overall communication, and if we lose one of those fragments, then we have to resend the entire frame again.
Normally the determination of the most optimal MTU or Maximum Transmission Unit is done automatically when you first connect to a device. But sometimes filtering and firewalls won’t allow the automated process to work properly. And you may need to manually configure an MTU on your system.
On some networks, we’re not trying to minimize the size of our communication. Instead, we’re trying to maximize the size of our frames. And one of the ways we can do that is through a method called jumbo frames.
Ethernet frames by default are 1500 bytes in length. If we’re transferring large amounts of data, we may want to increase the size of those frames to make the communication that much more efficient. And with jumbo frames, you can increase the size of an ethernet frame up to 9,216 bytes of an MTU.
On many devices. You’ll find that they’ll use 9,000 as the accepted high end of a jumbo frame. In fact, here is the configuration from my computer that says that I can configure frame sizes on my network to be between 1,280 and 9,000 bytes. This means I can transfer a single frame that contains the same amount of information as six normal frames on the same network. This means there are fewer packets to have to switch or route through my network and I can transfer much more in a shorter amount of time.
The challenge, though, is all of the devices that are in that communications path must understand jumbo frames. So all of your switches, all of your routers, and the other devices you’re communicating to have to also be configured to use these jumbo frames. If there is any device in the middle of this communication that doesn’t understand jumbo frames, then they will drop these frames and you won’t be able to communicate to that device at all.