Why we use collision domain?
I hope all is well. Let me start by giving you a definition of what a collision is:
A collision is described as an event that usually happens on an Ethernet network when we use "Shared Media" to connect devices. Shared Media is a type of connecting media which is used to connect different network devices, where every device shares the same media. Examples include:
- Ethernet Hubs
- Bus Topology
When we use Shared Media, there are not separate channels for sending and receiving of data, so everything uses one channel to send and receive the data signals.
Collisions happen when two devices simultaneously try to send data on the Shared Media, since Shared Media is half-duplex and sending and receiving is not supported at same time.
We use CSMA/CD to help us to deal with this by detecting that a collision has occurred and resending the dataframes that were lost.
Take a look at the following, as they will give you some additional background on the history as well as some pictures that will help you to make sense of everything:
Good Luck !!
Let me add to @Adam-Gordon 's excellent response. We need to use collision domains on shared network segment so that everyone has a chance to access the shared network.
The problem is that the more collisions we have the less efficient the network becomes. So the goal is to reduce the collision domains as much as possible by removing devices on that shared segment. Ironically, without reducing the number of host connecting, we do so by creating MORE but SMALLER collision domains with fewer hosts--using
bridges. This ultimately leads to micro-segmentation (only one device per segment). We do this by changing the central connectivity device from an ethernet
hubto an ethernet