IN THIS TUTORIAL
In this tutorial, you will learn:
- Meanings of words/phrases related to networking.
- Common network topologies.
- The Layers of the OSI model and a general summary.
In this tutorial, you will not learn:
COMMON JARGON DEFINITIONS
- how to set up/maintain a network.
- OS/kernel specific network functions.
- Details on OSI/DoD model or any protocols and apps.
- LAN - Local Area Network - network commonly spread across a distance of one building.
- MAN - Metropolitan Area Network - network commonly spread across a distance of around 35 miles or less.
- WAN - Wide Area Network - network spread across a distance of more than 35 miles.
- topology - The physical layout of a network.
- Node - Any device connected to a network (workstations, printers, servers).
- Server - Node that allows many other nodes to access its information.
- host - node (Typically workstation or server) that holds the data and sends the data.
- Client - node (Typically workstation) that recieves the data.
- Protocols - Sets of rules and guidelines which lay out the way in which data is formatted, prepared, transferred and recieved.
A topology is the physical way in which a network is set up. This means the format in which the different nodes are connected and by which means. For this part, I'll be using certain images from google to demonstrate. No credit should go to me for these pictures.
We need topologies to keep our networks in order, they form the base of our connections and which topology we use is highly dependant on how much money they cost, how efficient they are, how fast they are and the security/failsafes of them. Here are some examples:
This topology is pretty simple to understand. There is a cable, and on each end of the cable are terminators. These stop any data from reversing when they hit the end of the cable and doubling back. Each node is connected to the linear cable. If any data has to be passed, it must be passed along this connection. This layout is very uncommon now, as better ones have been found. For instance, consider the computer on the far left wants to send some data to the computer on the far right. To get there, the data has to pass through all the nodes on the way. The node then has to recieve the data and think "Is this for me?". If it is'nt, it re-transmits the data on to the next node. This causes a lot of overhead and is a slow process. There is also a possibility of collision, as all of these nodes are connected to one collision domain. This means that if one node sends some data in one direction, and another sends some in the other direction, the data can clash and end up not being recieved by the applicable node. Also, if the BUS topology has no collision-detecting mechanism, both nodes will assume their data has been sent successfully and the client has recieved the files. This can be prevented by using the CSMA/CD protocol, but that may be for another tutorial.
This topology is one where all the nodes are interconnected to one another. All connections lead from one node to another, and in a network of more than 2 nodes, there are at least 2 connections stemming from each node. A partial-mesh topology is one where some nodes are connected to all others, whilst other nodes are only connected to a select few. This topology is at an advantage in some aspects. For instance, if one node were to break or go offline, the network would still run and other routes would be available. This works well with small LANs, but problems arise when the network expands. The network becomes expensive to implement and I'm sure you can imagine how complicated the layout would become. There would also still be the problem of collisions.
The star layout is used much more than the above. As you can see, nodes are all positioned around a central node, which is typically some routing device that separates collision domains (i.e. router, switch, bridge). This central node handles the passing of datagrams (data packets) between other nodes. This topology creates one broadcast domain. This is because any data sent to the central node (let's call it a switch for now) is then passed on by the switch to every node it connects to. Therefore, if somebody broadcasts something, everybody on that network recieves it, thus there is one broadcast domain. This topology, though aimed at small LANs has some advantages. It's easy to implement and sustain and it's also easy to troubleshoot. If one node goes offline, the switch will just discontinue communication with that device and the network will function as normal. The switch also creates one collision domain for each node. However, a disadvantage is that if the switch were to fail, the network would cease to work. Also, if a lot of data were being transmitted on the network, the switch would become slow and thus not be very good for more than the simplest tasks.
TUTORIAL STILL IN PROGRESS