Week 4: Day 013 - Ethernet Basics #2

Hello to all readers, today we will be going back to the section on Ethernet. If you look back I did a detour to Ethernet basics hoping to learn about the cable enough to be able to create one myself. Thankfully, I was successful in creating an Ethernet cable after confirming it during a test! Maybe at one point I'll make a special post on how it worked out. Anyways, without further a due, let's get back to talking about Ethernet!

Last time, I left off talking about Frame Check Sequences (FCS) and Cyclic Redundancy Check (CRC). To put it in shortly, the CRC basically is a way of making sure your data arrives to it's destination intact and without any corruptions. If the CRC deems that the data didn't arrive intact it requests the sending computer to send it again. Basically what Network Engineers managed by implementing stuff like this and giving computers MAC Addresses as identifiers. However the problem of determining the timing in which a computer gets to send their data over the network requires something else. This solution is known as "carrier sense multiple access/collision detection (CSMA/CD) which is a way of determining which computer gets to use the cable, and when. "Carrier sense" means each node examines the cable before sending data frames. During these checks by the nodes, it waits until the cable is "free" to then send it's data. Now "multiple access" means that all computers and devices have equality over their access to the cable. From the Ethernet cable's perspective, all it knows is that it's own use is first come, first serve. It has no sense of what is important and what isn't (ex: windows/linux server has same access as an old XP.) So what happens when two machines decide to send data at the same time? What happens is called a "collision". When this occurs the data is completely destroyed, and needs to be sent again. It's pretty easy for NIC's to detect collisions, and when they do, the nodes stop sending data immediately. Funny enough, the two machines will randomly generate a number time and send their data at that time, just like rolling dice. The one that wins gets their data on the wire first. Collision rates are 10% on a normal Ethernet network, and are completely normal.

Next, we'll talk about Early Ethernet Networks. Back in the day, the things that I mentioned earlier in the entry did not exist, and were nuisances to those who administered networks. Anyways, let's talk about the different types of cables that they had early on. First, there's the 10BaseT which was created in 1990 by the IEEE 802.3 committee, and this modernized the Ethernet cable. Quickly after release it became extremely popular. In modern day, 99% of networks use 10BaseT and its newer/similar versions. The classic network pertaining to this cable has 2+ computers connected to a central hub. The NIC's connect through wires. The hubs used by 10BaseT come in many shapes and sizes. The differences in the hubs are the number of ports they have. Small ones only have 4+ ports, while big ones have 48+ ports! Each hub needs power, if it doesn't get enough, the entire segment will cease to work. The naming of 10BaseT refers to it's speed of 10 Mbps, it's signaling type: baseband, and T for the fact that it uses twisted pair cables. Speaking about UTP cables, let's move on to that subject! The 10BaseT hubs have a prerequisite of CAT 3+ and two-paired UTP's. Usually installers would four-pair them for insurance against higher requirements needed for future networks. They are equipped with Kevlar fibers to stop you from destroying the cable. The 10BaseT also added RJ-45 connectors to the networking world. The 10BaseT gives wires numbers to differentiate their purposes. However, a 10BaseT device cannot send and receive data at the same instance, unfortunately. This is because CSMA/CD rules apply here. Now I think I'll skip the part about crimping and how to create a cable, because I've already done it myself, and may post something about it later. So to summarize this, the summary of the 10BaseT based on my textbook is: Speed - 10 Mbps, Signal Type - Baseband, Distance - 100 meters between hub and node, Node limit - Limit of 1024 nodes per hub, Topology - Star-bus topology (Physical: Star, Logical: Bus), Cable Type - CAT 3+ or UTP cabling w/ RJ-45 connectors. A few years later they released the 10BaseFL which was a fiber-optic version of the T. With all the benefits of fiber-optic cabling, this was a good choice for environments that had the option of a Fiber-optic network. The 10BaseFL uses multimode fiber-optic and needs an SC or ST connector. The specifications of the 10BaseFL is: Speed - 10 Mbps, Signal Type - Baseband, Distance - 2000 meters between hub and node (much bigger!), Node Limit - Limit of 1024 nodes per hub, Topology - Star-bus topology, Cable Type - Multimode fiber-optic with ST or SC connectors. So those are the differences between both the 10BaseT and 10BaseFL. It should be noted that a media converter can be used to interconnect both types. On that note, I will end this blog entry --- another long one! See you next time, tschuss!