CSMA/CD vs CSMA/CA Collision Detection

[Mitrixsen]

Hello, everyone.

I have some questions as to what the difference between CSMA/CD and CSMA/CA is.

First of all, how exactly do these two differ? From my understanding after reading the explanations to them, CSMA/CD is capable of detecting and recovering from collisions after transmission has begun while CSMA/CA cannot do that therefore it aims to avoid them completely?

So to help me imagine this further, why cannot wireless clients also listen for collisions, detect them, and recover once transmission has begun? The explanation is often that there is no cable like in wired networks but is it really not possible to detect whether there are two wireless signals that bumped into eachother?

And the final question is, I understand that the RTS/CTS process is involved in the background as a part of CSMA/CA. My question is regarding the following paragraph of information:

After successfully transmitting data, the sending device awaits an acknowledgment (ACK) from the receiving device. If the ACK is not received within a specified time, indicating that the transmission may have been unsuccessful (possibly due to a rare collision or other issues), the sending device will attempt to resend the data after waiting for a random backoff period.

What exactly is this acknowledgement from the receiving device? What if the data eventually travels over a wired network to the receiver? How would the receiver know that it needs to send an ACK back?

That's all, thank you!
David

[marce1000 · Accepted solution]

            -  FYI : https://www.pynetlabs.com/csma-cd-vs-csma-ca/

    M.

[Rich R · Accepted solution]

The link @marce1000 shared provides a very good concise explanation so have a good read through that.

> What exactly is this acknowledgement from the receiving device? What if the data eventually travels over a wired network to the receiver? How would the receiver know that it needs to send an ACK back?
You're confusing your OSI layers - this is layer 2 (data link) of the OSI stack so the ACK is between the wireless devices which are sending and receiving the frame (IEEE 802.11 protocols).  It's irrelevant if the frame is transmitted on some other medium after the wireless device has received it because it's the wireless device which is ACKing receipt of the frame, not the end destination of that frame/packet. The "receiver" is always a wireless device, though the data in that frame might ultimately be destined to be forwarded on to somewhere else.

@Rasika Nayanajith provides some packet capture illustrations at https://mrncciew.com/2014/10/26/cwap-802-11-ctrl-rtscts/

[Scott Fella]

I remember when I started long ago, I asked this question to folks also, but this was a while ago when companies still had ethernet hubs not switches and or a mix.  They were also implementing wireless 802.11b back then.  What a peer explained to me and I will summarize, is that CD is built for ethernet and CA is built for wireless.  The issue with wireless is that is is a half duplex system as a hub would be, but in wireless, when clients are trying to send its traffic to an access point which causes utilization on that channel that he ap and device is using.  That is why it makes sense to try to AVOID vs DETECT when it cones to wireless. That why they have these backoff timers to help reduce retransmission on the wireless.  Just keep in mind that the CD and CA are just from the device to the hub/switch/ap nothing behind that.  AP to switch would use CD.