(You Got Your Cabling In My
Ethernet!)
When we get to Wide Area Networks, we’ll run into quite a few different encapsulations -- HDLC, Frame Relay, and PPP in particular.
With Local Area Networks,
whether we’re connecting a host to a switch….
… or we’re connecting two switches…
… or we’re connecting a switch to a router…
…. we’re likely using Ethernet and Ethernet cables. In this section, we’ll talk a bit about how Ethernet works, and the different types of cables we’ll use in this network. And note that I said “Cables With
An S”, because not all Ethernet
cables are the same!
Actually, not all Ethernet types are the same, so let’s start there.
Not All Ethernets Are The Same
“Ethernet” is really an umbrella term at this point, encompassing several different types of Ethernet, different capacities, and different challenges. For both a successful exam experience and a solid
networking career, it’s a great idea to be comfortable with these values.
Most kinds of Ethernet cables are Unshielded Twisted-Pair (UTP). The name is the recipe – the wires inside the cable are indeed pairs of
twisted cables.
Why twist ’em? Twisting pairs of wires inside the cable cuts down on electromagnetic interference (EMI). EMI can interfere with the
electrical signals carried by the wires, which in turn is really going to screw around with our network.
EMI can come from other cables, and also (and infamously) from elevators. I know of more than one network that would slow down at lunchtime and quitting time because
that’s when the elevators were in heavy usage, and the network cables werun right next to the elevator shaft, which in turn gave our network the shaft.
We can even have EMI problems from other wires in the same cable! This crosstalk happens when a signal “crosses over” from one pair of wires to another, making the signal on both sets of wires unusable.
when wires are crossed or crushed. The conductors inside the wires don’t have to be exposed, but if the conductors are too close, the signal traveling on one wire can interfere with the signal on another wire.
Here are some common Ethernet types that run on regular old copper cabling, along with their official IEEE name and more common name. All have a maximum cable length of 100 meters.
802.3, is generally referred to as 10Base-T, and runs at 10 Megabits per second (Mbps).
Fast Ethernet (802.3u) is usually called 100Base-T, and runs at 100 Mbps.
Gigabit Ethernet (802.3ab) is generally called 1000Base-T, and runs at 1000 Mbps.
NOT called 10000Base-T. It’s usually called 10GBase- T, and runs at – you guessed it! – 10 GBPS.
There’s a huge difference between 10GBase-T and 10Base-T. Watch that G!
We also have a version of Gig Ethernet that runs on fiber-optic cable. That version is 802.3z, and is often called 1000Base-LX. This version has a max cable length of 5000 meters, as opposed to 100
meters with all the other versions we’ve seen.
So with that huge max cable length, why aren’t we running everything on 802.3z? It’s the sheer cost of the fiber optic cable. It’s a lot more expensive to install and
Multiple Standards Usually Equal Multiple Nightmares
Luckily, this isn’t one of those situations.
When you send an Ethernet frame from Point A to Point B, there’s a chance the frame could go across a “regular” Ethernet link, then a Gig Ethernet link, and then a Fast Ethernet link as it arrives at its destination.
If we had to do some kind of
translation every time a frame went from one Ethernet type to the other, we’d be doing a lot of translations and adding big time to our
transmission time and overall network workload. Fear not – we don’t have to do anything like that. All of our different Ethernet
standards have the same overall frame format:
There’s another tool that allows us to seamlessly use network and host devices with different Ethernet capabilities – autonegotiation.
Autonegotiation didn’t work all that well years ago, and it got to the point where most network admins manually set card and port settings. Cisco went so far as to make it a best practice NOT to use
autonegotiation.
I mention this because I don’t want you more-experienced network
admins glossing over this section, thinking “Hey, autonegotiation doesn’t work.” Autonegotiation has come so far since the bad old days that it’s actually mandatory for Gig Ethernet over copper, and it’s an important part of the overall Gig Ethernet standards.
Having said that, let’s see how autonegotiation works!
In this example, we have a host device connected to a Cisco switch port. The host is running 10BaseT,
and the switch port has a top capability of 1000BaseT.
The devices announce their capabilities via Fast Link Pulse (FLP). The logical question: “Fast as compared to what?”
Fast as compared to the Normal Link Pulse (NLP)! Basically, the NLP is sent by an Ethernet device
when it has no data frames to send – it’s saying “Hello, I’m still here!”
Here’s the NLP, compliments of Wikipedia:
Here’s the Fast Link Pulse, which autonegotiation-enabled devices use to announce their capabilities:
After the devices exchange this information…
… they come to an agreement on the values to use. As you’d expect, the
lowest speed is the one selected, and full-duplex is preferred over half-duplex. In this situation, the PC port and the switch port it’s
connected to would run at 10 Mbps, full duplex.
Autonegotiation will dynamically adjust if a port capacity changes. Let’s say you replace that PC with a PC that can run at Fast Ethernet speed.
If we manually set all of our switch port settings, we’d have to change the speed on that port manually. With autonegotiation, the switch will realize the new capacity of the device connected to that port, and voila – a 100 Mbps link!
I highly recommend you use autonegotiation on both ends of a link such as this one, or don’t use it at all. You can end up with a link that isn’t working at its real
capacity due to a duplex mismatch – a link where one endpoint is running at half-duplex and the other
end is running at full-duplex.
When using Cisco switches, if autonegotiation is turned off on the other end of the link, the switch should still be able to sense the speed capacity of the other endpoint. If for some reason the speed capacity can’t be detected, the lowest speed supported will be used.
That probably doesn’t surprise you, but this might. If that detected speed is less than or equal to 100 Mbps,
the switch will set its port speed to half-duplex. Hello, duplex
mismatch!
In this example, the Cisco switch has successfully detected the capacity of the remote endpoint to be 100 Mbps. No problem there, but a problem does arise when the Cisco switch sets its port connected to that host to half-duplex as a result of that speed.
Duplex mismatches have a special place in Network Heck, because they can be difficult to spot. The two devices will still be able to exchange data, but it’s going to be a slow, inefficient process.
Run autonegotiation on both ends or don’t run it at all.
Crossover and Straight-Through Cables
We’re going to use a simple
network for this demo, and the two separate physical connections will require two different cable types.
2 and a switch sends on pins 3 and 6. In turn, the PC receives on pins 3 and 6, and the switch receives on pins 1 and 2. This means we can use a straight-through cable to connect the PC to the switch.
The cable name comes from the wires inside the cable. Assuming we’re using Ethernet or Fast
Ethernet for this connection (a safe assumption), we’re going to have four wires inside the cable, and each wire goes straight through from one end to another.
What exactly does “straight
through” mean in this situation? The wire connected to Pin 1 at one end goes straight through to Pin 1 at the other end, the wire on Pin 2 goes straight through to Pin 2 at the other end, and the wires on Pins 3 and 6 go straight through to those pins on the other end.
If you enjoy making your own cables and you run into a
connection issue right away, I can practically guarantee the problem is that one of those wires in your
over to another pin.
Gigabit Ethernet can use straight- through cables as well, but to carry data that quickly, it follows that we’ll have more wires inside the cable. Where Ethernet and Fast Ethernet have 4 overall wires inside the cable, Gigabit Ethernet has eight. In a Gigabit straight- through cable, one wire goes from Pin 1 to Pin 1, one wire from Pin 2 to Pin 2, and so forth for all eight pins.
Wires crossing over inside the cable isn’t always bad. Sometimes
we want those wires to cross over in the cable – hence the name “crossover cable”, our next cable type!
Crossover cables are necessary when we’re connecting two devices of the same type, and in a typical network, that’s going to be two switches. When we tackle
switching in this course, you’ll see why interconnecting switches is so common. Our first step in this
interconnection is choosing the right cable!
We can’t use a straight-through cable for a switch-to-switch
connection, since they use the same pins to send and receive. We’d have the same pins sending data on both ends (a bad idea) and pins 3 and 6 on each end listening for data that will never arrive (sad!).
Communication between the switches is made possible with a
crossover cable. The four wires
inside the cable “cross over” from one pin to another inside an
Ethernet or Fast Ethernet crossover cable:
Wire on Pin 1 crosses over to Pin 3
Wire on Pin 2 crosses over to Pin 6
Wire on Pin 3 crosses over to Pin 1
Wire on Pin 6 crosses over to Pin 2
With this setup, when a switch sends data on the two pins used to send data (Pins 1 and 2), the switch on the other end of the cable will receive the data on pins used to receive data (Pins 3 and 6).
Gigabit Ethernet crossover cables have those same wires cross over in addition to the following:
Wire on Pin 4 crosses over to Pin 7
Wire on Pin 5 crosses over to Pin 8
Wire on Pin 7 crosses over to Pin 4
Wire on Pin 8 crosses over to Pin 5
Now it’s time for a little “real world vs. theory” chat.
After reading that cabling section, some of you are saying “Hey, I used a straight-through cable to connect two switches with no trouble.” And you’re right – you just might have. Most Cisco switches will recognize what you’re trying to do when you connect them to each other with a straight-through cable, and the switch will dynamically adjust itself to make the straight-through cable work.
When it comes to your CCENT and CCNA tests, though, you need to forget about that. Be clear on when you’d use a straight-through cable as opposed to a crossover cable:
Devices transmit on same pins = crossover cable
Devices transmit on different pins = straight-through cable
Both straight-through and crossover cables end with RJ-45 connectors, which “snap” right into place when connected to a PC NIC or switch / router Ethernet port.
Several protocols and services you’ll be introduced to in this course have more than one name, and we’ll start that tradition with the next topic, known by all of these names:
MAC address (Media Access Control)
Physical address (because the address physically exists on the network card)
Burned-In address (BIA – the name comes from the address being physically burned into the NIC)
Ethernet address
That’s nothing! We used to have
seven names for this address, but
the terms “NIC address” and “LAN address” have pretty much fallen by the wayside. Throughout the
courses, I’ll use the term “MAC address”, but you should be familiar with all the names listed
here.
The MAC address is used by switches to send frames to the proper destination in the most
efficient manner possible, a process you’ll be introduced to in the
Switching section. Before we see how that works, I want to introduce you to the address format and the characters we’ll see in this address.
The MAC address is six bytes long (48 bits), and can be expressed in either of these formats:
aa-bb-cc-11-22-33 aabb.cc11.2233
MAC addresses consist of hexadecimal values, and if that phrase gives you anxiety,
fuggetaboutit. By the time this section is over and you get some practice in, you’ll be working with hex like a champ – or more
accurately, like a CCENT and CCNA!
You’ll hear me say this throughout the course, and I’ll start now: The
key to mastering hex, binary, and subnetting is practice. Reading about it is not enough!
The MAC address has two parts, the first being the Organizationally Unique Identifier (OUI, not
pronounced like the French word for “yes”, but Oh-You-Eye). The OUI is assigned to hardware vendors by the Institute of Electrical and Electronics Engineers (IEEE).
unique to that organization and is not assigned to another org.
The second half of the MAC address is simply a value not previously used by the hardware vendor with that particular OUI. Using the earlier MAC address example, we see that…
The OUI of the address is aa- bb-cc
11-22-33 with that particular OUI, so the vendor is doing so now
There’s a special MAC address for broadcast frames, and as we get to that topic, let’s take a look at the three overalltypes of network traffic.
Unicast traffic – Destined for one particular host
for a group of hosts
Broadcast traffic – Destined for everybody
You can spot broadcast and
multicast MAC addresses by using the following rules:
The broadcast MAC address is ff-ff-ff-ff-ff-ff (or FF-FF- FF-FF-FF-FF, as case doesn’t matter in hex).
We have a range of multicast MAC addresses. The first half of a multicast MAC address will always be 01- 00-5e. The second half will fall in the range of 7F-FF-FF. Watch that 7!
Remember that Ethernet header and trailer I mentioned briefly?
No?
Well, I don’t blame you, it was a fast mention. Let’s take a more detailed look at both the header and trailer.
The Ethernet Header And Trailer
Here’s a high-level look at the overall Ethernet frame:
A detailed look at the header:
From left to right, a quick look at each field:
The preamble is there for synchronization purposes. The nuts and bolts of this field are (thankfully) way beyond the scope of the
CCENT and CCNA exams. If you’d like to read more about this field, check out the
Wikipedia entry for Ethernet.
The Start Frame Delimiter (SFD) indicates the preamble has ended and the destination MAC address is on deck.
Both the destination and source addresses are MAC addresses.
Finally, the type (EtherType) field indicates the protocol type carried in the data field. In today’s networks, that’s likely IPv4 or IPv6, but it can be plenty of other protocols.
Here’s a detailed look at the Ethernet trailer contents:
That’s it!
Considering the FCS is the Ethernet caboose, it’s easy to think there’s not much going on there, but the FCS is a vital error detection tool. It’s basically a three-step process:
The sender runs an algorithm against the contents of the frame, and takes the result of that algorithm and puts it in the FCS field. The result is the checksum.
The receiver runs the exact same algorithm against the same contents, and expects to come up with the same
checksum contained in the FCS field of the incoming frame.
If the results are the same, the frame is fine. If the results are not the same, something happened to the frame contents as the frame went across the wire, and the frame is dumped.
There is no explicit notification from the receiver to sender that the frame was discarded. The FCS brings us error detection, but not
correction.
We saw two common network connections earlier (PC to switch, switch to switch), and there’s
another one I want to introduce you to – connecting your laptop or PC directly to a switch or router in order to configure it. For that
physical connection, you’ll need yet another type of cable.
When you physically connect your laptop to a router or switch, you’ll
be connecting to the Console port on the network device. For this, you’ll need a rollover cable, also called a console cable. There are 8 wires inside the rollovercable, and they each roll over to a different pin at the other end: Pin 1 to Pin 8, Pin 2 to Pin 7, Pin 3 to Pin 6, and so on.
One end of the console cable will have an RJ-45 connector, similar to that on the end of a land line phone wire. You’ll feel (and maybe hear) that end of the cable snap into the Console port.
It’s the other end of the console cable you need to be aware of. Some console cables have a DB-9 connector on one end, and modern laptops don’t have a DB-9 port. If that’s your situation, get an adapter for your cable – you can find them online at any major cable dealer. (And even most minor ones!)
The need for an adapter is a good thing to find out before you visit a client site.
since they’re almost totally flat and usually colored light blue.
Here’s a link to a page on
PacketByte.com that shows you a console cable, along with two other cable types that you’ll find in any great Cisco home lab!
http://packetbyte.com/Content/Cabling/OtherCables.html
Next up – hubs and repeaters. You might not see many of them in today’s networks, but you need to
understand how they work in order to really grasp switch operation – and you’ll see hubs, repeaters, and switches on your exam, so let’s hit it!