Introduction to Ethernet

Lessons Discussion
#4

Hello Mohammad

Yes, there are several types of Ethernet framing, two of which are the most common that you mention above.

Ethernet II has a two byte EtherType field that identifies the upper layer protocol (e.g. IP) that encapsulates the frame. For example, a value of 0x0800 indicates an IPv4 packet while a value of 0x0806 indicates an ARP frame.

The IEEE 802.3 standard changes the role of the EtherType field to a “Data Length Field”.

Later on, because both types were in wide use, IEEE 802.3 incorporated Ethernet II and standardised both formats.

Today, both of these formats can coexist on the same network. The method by which network devices know how to interpret each frame as one or the other is by the EtherType /Length field. If it is >= 1536 (0x0600) then it is an Ethernet II frame and that field is interpreted as an EtherType field. If it is <= 1500 it is an 802.3 frame and that field is interpreted as a Length field. Values between 1501 and 1535 are considered undefined and under normal circumstances should never be encountered. Therefore both standards can use the same medium.

I hope this has been helpful!

Laz

Ethernet Introduction for CCNA R&S
#5

Hi Rene,
can you explain about auto-mdix feature.

thanks
aditya

#6

Hello Aditya

Traditionally, routers, switches, and computers require a different type of Ethernet cable to connect to each other. There are crossover cables and straight through cables, as described in the lesson.

What auto MDIX does is it automatically detects the kind of cable required for a particular connection, and if the cable used is the wrong type, then it automatically switches the appropriate pins internally on the switch or router, such that the correct connectivity is achieved. So for example, if you use a crossover cable to connect a PC to a switch (which is the incorrect cable type) the auto MDIX feature will detect this and internally switch the Tx and Rx pins so that the connectivity is achieved.

I hope this has been helpful!

Laz

#7

Hi Rene,
Firstly, I am enjoying your lessons and tutorials. They are very informative and clearly laid out.

Nonetheless, I have been doing some thinking and trying to understand how an Ethernet Frame is encapsulated then decapsulated through the layers. So I have a question about the method an ethernet frame if transported from say Host A to Host B via a SOHO home router/modem.

Let’s say Host A sends an Ethernet Frame over the Internet to Host B. Host B is on a private LAN network with some other devices. The Ethernet Frame arrives at Host B’s SOHO home router/modem and for argument sake the frame has passed through the modem part and NAT has translated the destination IP address from public address to private address. This is where I am a little confused and have a question: What happens next?

I am assuming it goes to the router part of the SOHO device (layer 3). But the Ethernet Frame is layer 2 and only shows the destination MAC address. So, does the Ethernet Frame get decapsulated at this point to reveal the IP Packet for the router to understand the destination IP address for a split moment to then be encapsulated again when it goes to the switch part of the SOHO router/modem for it to be routed to Host B. Or does the Ethernet Frame decapsulated to layer 3 and the IP packet is sent to the switch where an ARP broadcast request is sent to locate Host B?

I have understood the lessons on ARP and the Ethernet Frame and IP Packet with the simple examples of a direct link. But not sure how it functions when an obstacle like a router is put in between. It seems there are more processes than needed to get a frame from A to B.

I maybe over complicating my thoughts or running ahead of myself here. But it is something I have been pondering about which I can’t get out of my head and need a clearer understanding.

Adrian G

#8

Hi Lazaros,

I have a question on Ethernet Frame size that I am a little confused about now reading your answer to Mohammad’s question. The Introduction to Ethernet lesson explains that the Ethernet Fame is made up of 7 components (in its basic form), Preamble (7 Bytes), SFD (1 Byte), Destination MAC (6 Bytes), Source MAC (6 Bytes), Type (2 Bytes), FCS (4 Bytes) then Data (46 - 1500 Bytes). If we add all these component fixed sizes (excluding Data) we get 26 Bytes of overheads. Therefore adding this overhead total to a maximum possible data payload of 1500 Bytes, we get 1526 Bytes. I have presumed Rene was referring to an Ethernet I Frame in the lesson. So why have you stated anything less than 1500 would be Ethernet I Frames and any Frame larger than 1536 is Ethernet II Frames with anything in between considered undefined ?

I have tried to research this online but everywhere concludes to the same answer with little explanation. It seems to me that an Ethernet I Frame could be larger than 1500 Bytes upto 1526 Bytes in any normal situation so not sure why it will be considered in the undefined range.

Can you explain further, or have I got this all wrong?

Adrian G

#9

Hello Adrian, good to have you with us! Glad to hear that you’re enjoying the lessons!

In order to understand what is going on, let’s take it step by step. The packet has arrived from the Internet at the SOHO router. Now we don’t know what kind of header we’ll find at layer 2, as this depends on the technology being used. If it’s ADSL, serial link, or cable modem, each of these have a different L2 technology, and therefore a different L2 header. Whatever it is, the L2 header is stripped off, and we are at Layer 3 of the OSI model. So we take a look at the IP header. It has the outside (public) address of the router as a destination address.

Now is you correctly stated, this IP address is translated using NAT, and the corresponding inside IP address replaces the destination address. Once the IP header is correctly translated, the next step is for the SOHO router to decide where to send the packet. In most cases, there is only one choice, as there is only a single routed port on the inside of such devices. (Remember that a SOHO router connects the WAN to the LAN using layer 3 routing).

So the router then has to add a L2 header with the appropriate MAC address. In order to do so, it takes a look at its ARP table to see if this destination address (the newly translated internal address) exists in its ARP table. If it does, the destination MAC is populated, the source MAC is it’s own, and the frame is sent. If it doesn’t exist in the ARP table, then it sends out an ARP request for the particular destination IP, receives the destination MAC address, populates the appropriate field in the L2 header and sends it on its way.

I hope this has been helpful!

Laz

#10

Hello again Adrian

The method by which network devices know how to interpret each frame as one or the other is by the EtherType/Length field. We’re not talking about the actual length of the frame itself, but about the numerical value found within that field. If the numerical value is 1500 or less, then it is interpreted as an IEEE 802.3 Ethernet frame, and the numerical value is interpreted as the size of the payload in bytes. If the value is 1536 or more, then the frame is interpreted as an Ethernet II frame, and thus the numerical value is used to indicate what protocol is encapsulated in the payload of the frame. For example, a value of 2048 refers to IPv4, 34525 to IPv6 and 34887 to MPLS unicast. Note that all these values are greater than 1536 and thus are interpreted as Ethertype and not length.

I hope this has been helpful!

Laz

#11

Hi Laz,

Thank you for your time replying to my question. Your step by step method answer made it much clearer and I can see where I was a little confused now from the Layer 2 to Layer 3 to Layer 2 transition.

So that I am clear and have understood: From the internet (cloud / public network) the layer 2 frame comes in different forms, which maybe ADSL, serial link and cable modem. These are not pure ethernet frames!!! so when the frame arrives at a SOHO router this layer is naturally stripped off to reveal the layer 3 IP packet for NAT translation. This IP packet will show a public destination address that the NAT translates by removing the layer 3 to reveal a layer 4 destination port number. This port number is checked on the NAT table to identify a private (LAN) IP address associated with the port number. A layer 3 header encapsulates the layer 4 segment with a new destination IP address.

The router looks up its ARP table to identify a MAC address associated with the new destination (private) IP address. A layer 2 header encapsulates the layer 3 packet with a destination MAC address. The layer 2 ethernet frame is sent on its way across the LAN network to locate the destination host.

Adrian G

#12

Hi Laz,

Thank you for your reply. As always explained very well to the question.

I see that I have miss understood between the size of the Ethernet frame and a numerical value in the EtherType/Length field. Thank you for clarification and I will read up a bit more on this subject. You have been very helpful, thank you.

Adrian G

#13

Hello Adrian

Your explanation is correct, that’s how it works. Now jut to clarify, the Layer 2 technology used to reach your home/office will determine what exists at Layer 2. But that exists only for the portion of the transmission between your ISP and your premises. However, layer 3 remains unchanged throughout the whole transmission from end to end (except of course where NAT is involved). This is the beauty of the OSI layered model, you can change Layer 2 technologies without touching the functionality of Layer 3 and above.

I’m glad this has been helpful!

Laz

1 Like
#14

I am using this website as a review after going through the “How to Master the CCNA” text. There are a few differences concerning the ethernet frame when comparing this site to the text.
Network Lessons vs How to Master…
SDF vs SOF
Type vs Length

Also, Data is 802.2 Header/Data per the text. If the encapsulation mode is ISL, would it still have this title?
Are the differences listed above relevant for the exam(s)?

Thanks

#15

One more comparison of the text to this site.

The MAC address has a BC and LC field that are 2 bits combined. They are not present per the Intro to Ethernet section on this site. Does it matter? If so, which to adhere to during the exam?

Thanks

#16

Hello Oliver

The differences that you are stating correspond to more detailed information that is available in the text. I’ll attempt to explain these differences.

  1. Concerning Type vs Length, Ethernet comes in two formats: Ethernet II and IEEE 802.3. In Ethernet II, the specific field is known as EtherType or Type, while in IEEE 802.3, the field is known as Length. In today’s networks, both Ethernet types are supported. Now if you use Ethernet II, the Type field will contain information that tells the device the type of upper layer protocol that is encapsulated. For example, 0x0800 means IPv4, 0x0806 means ARP, and 0x86DD means IPv6. However, if you are using IEEE802.3, then this field only contains information concerning the length of the data field. In such a case, the recipient still needs to know how to interpret the upper layer protocols. That’s where the 802.2 header comes in.
  2. 802.2 Header/Data field is used when the Ethernet being used is IEEE802.3. What this does is place an extra header called an 802.2 header as part of the data field. In that field, you will find the information concerning the type of upper layer protocol being used. The 802.2 standard operates on the Logical Link Control (LLC) sublayer of the Datalink Layer.

As for the SOF, SDF, BC, and LC fields, can you elaborate on those? I’m not sure I understand what you are referring to.

For the CCNA level exam (and probably for CCNP as well) these details are not pertinent. I believe it is enough to know that the two types of Ethernet exist, that one has a Type and the other has a or Length field, and to understand the purpose of the Logical Link Control (LLC) and Media Access ontrol (MAC) sublayers.

I hope this has been helpful!

Laz

#17

Hi

Thank you very much for thorough explanation. I benefited a lot from this lesson

1 Like
#18

Hi Guys - forgive me if I’m wrong but isn’t the OUI code 24 bits and not 22 as stated in the lesson?

#19

Hi Gareth,

You are right, just changed this.

The first byte has two bits that have another function:

  • The unicast/multicast bit
  • The globally unique / locally administered bit.

The standard defines the entire 24 bits as the OUI though, so that’s what it should be.

Rene

#20

Hi Rene ,
As you mentioned in your post that we can use crossover cable if the communicating devices on the end side are same . Would you please tell me in what case or scenerio we can use cross cable if the device on both end is router As i know router contains both serial port and Ethernet port , we generall might use serial cable . So my question can we use cross overcable connected to etherent port for wan or internet purpose .
please share your relevant suggestion as it create confusion .

thanks
shivam chahal

#21

Hello Shivam

A crossover cable is used whenever the devices to be connected are the same (or when connecting a PC directly to a router). But this is the case only for Ethernet connections using RJ-45 connectors and UTP cable. This is not the case for serial connections or fiber connections of any type.

The concept of crossover cables only applies to copper UTP cables with RJ-45 connectors on an Ethernet link.

So if you have routers on both ends, and you require RJ-45 connectors for Ethernet, then a crossover cable should be used. It doesn’t matter if you are connecting to the Internet or to some internal network.

I hope this has been helpful!

Laz

#22

Hello Shivam

You seem to have replied to me in a private message. I will post it here for the benefit of other readers as well. You responded:

Hi Lagapides

Thanks for info but I want to know it’s practical implementation of crossover cable connected between the router and confirm me is the ethernet port in router can be used for wan link to service provider .You said that same device at end have use cross over cable but this cabling is used for ethernet connection.This statment is contradict .Please share in more deep as I m beginner in networking and confused to associate cable in gns 3 or packet tracer.

Why the statement is contradict that we use cross cable if device at both end is router as I know router is used for wan purpose or to communicate to other network for providing internet connection by choosing best path from routing table
We only use serial cable at serial port for wan connection and ethernet port of router is used to connect lan part .by connected straight cable.
Please tell me in ds what a router behave if we use cross cable .

Thanks & Regards
Shivam Chahal

I understand your confusion. First of all, let’s take a look at the definition of a WAN. A WAN is a type of network that connects the local network in a building or campus to either the Internet, a private network connecting remote sites, or to a network maintained by an ISP. The technology used by that WAN can be anything. It can be Ethernet (Metro Ethernet), serial, fiber optic, microwave link, xDSL, Cable modem, ISDN, MPLS, or others. A WAN link can use many different types of technologies, but what makes it a WAN is the geographic scope. In general, a WAN will connect a local network to the outside world, or to remote sites.

Secondly, a router is not limited to functioning only as a device connected to a WAN. Routers can exist within the enterprise network of a building or campus, and route traffic between internal subnets.

So you can have routers that do not connect to the WAN as well as routers that connect to the WAN. In both cases, many different types of technologies can be used for connectivity. WANs are not limited to serial connections and LANs are not limited to UTP based Ethernet cables. So you may have a crossover cable connect two routers inside your local network, or you might have a router connected to the WAN using an a crossover cable as well. Or you may have fiber optics within your network interconnecting many routers or a fiber connection providing you connectivity to the Internet or to remote sites.

In the past, WAN meant serial, and LAN meant Ethernet. This is no longer the case as many of these technologies, especially Ethernet, have been developed for use on the WAN as well.

I hope this has been helpful!

Laz

1 Like
#23

A post was merged into an existing topic: How does a switch learn MAC Addresses