If you read the last post, you know that I was trying to complete Chapter 2 and it just wasn’t happening. I realized this morning that I believe I have a cold which was causing me to not be able to keep going.
So, it’s probably a good thing that I broke this portion off as a separate section.
As a side note, I am much happier now with using WordPress. As I didn’t want to become a WP fanboy, it is the most popular CMS software and therefore much easier. I know it works, it works the way I need it too and I can change it if need be. Plus, there is documentation like a motherfucker.
I thought it was kind of weird that my fingers were starting to hurt last night.
anyway… back on topic.
So, we are going to start with Ethernet addressing. Ethernet addresses go by many names (Ethernet Address, MAC address, Physical address and Manufacturers address to name a few). But they all refer to the same thing. We generally use the term MAC Address (Media Access Control). The address itself is 6 bytes long (8 bits x 6 = 48 bit) that is notated in Hex. (which means 0 through 9 plus A through F). Normally you will see it with : between the bytes. Cisco does it differently to fuck with our heads. (I think so you know you are on a cisco device).
The first half of the address is the OUI or Organizationally Unique Identifier. This is a number that is issued to a manufacturer unique to that manufacturer. The second half of the MAC address is vendor assigned. No two MAC addresses are the same. This ensures that not only that each device on a LAN has a unique MAC address, but we can actually tell (somewhat) what type of device is connected.
Also it is important to note that this is considered at Unicast address. (The term unicast was chosen mainly for contrast with the terms broadcast, multicast and group addresses.) In the early days of LANs where everybody was connected on a bus or hub, they received all traffic. But to know if the traffic was meant for them, they would compare the MAC address listed in the frame to determine if it was for them or not.
For lan cards, they come with what is called a BIA (Burned-in Address) also called UAA (Universally Administered Address). But seriously, I think the author of this book just threw that in there. Cause like I said, everything in Networking is abbreviated.
There is also a group address which identifies more than one LAN interface card. The IEEE defines two general catagories. Broadcast and Multicast.
Broadcast is for all devices to receive. Normally FFFF.FFFF.FFFF
Multicast addresses are used to allow a subset of devices on a LAN to communicate. These follow the format of using 0100.5eXX.XXXX. These are used when one sender is sending to multiple receivers. Similarly to like a bunch of people receiving a video feed.
Enough on this crap, lets move on to Ethernet Framing.
First…we need a big hammer.
Wait, that’s house framing. We are doing Ethernet Framing. But the idea is not too different from the framing of a house. As stated before the Ethernet frame is a wrapper around the data from the adjacent layer. Which with layer 2 would be the packet from layer 3.
It also refers to the definition of the fields of the wrapper.
One of the more important fields is called the “Type Field”. Since there are many layer 3 protocols available, it is important to know what type or protocol is in use. So in one of the changes that were made in the Ethernet Framing protocol was to split the Length field to include the type field. (Similarly to DIX which used the type field instead of length. The preamble would be 8 for DIX and 7 for 802.3).
To produce the split, the length field had to be less that 1536 decimal. Then the type was broken into two, the LLC and the SNAP. LLC (Logical Link Control header) and the SNAP (SubNetwork Access Protocol header).
Finally, error detection.
With all the crap we are trying to pound into the frame, there are bound to be some errors.
So from the beginning it was determined that some error detection would need to be in order.
To do so, they used the trailer portion of the frame known as the FCS (Frame Check Sequence).
*seems legit*
Basically the frame has a calculation performed on it by the sender and added to the end. The receiver does the same math and if it does not match the trailer of the packet, it’s tossed out. (not tossed off…tossed out!)
There is no error recovery or anything done to resend the frame. Layer 4 protocols like TCP exist to handle that kind of action. Which we will go into a few chapters later.
But that is it!
That is finally it. Which is good because this took way longer to figure out how I was going to write the fucker than I should have.
But keep in mind that I think I had a cold through most of the last one.
****
Next in the pain that is this ongoing series of Network learnings…WANs! Yeah!