Talk:Carrier-sense multiple access
|WikiProject Computing / Networking||(Rated C-class, Low-importance)|
Restructuring of CMSA, CMSA/CD, CMSA/CA Pages (1 Nov 2003)
Need to add: Deterministic CSMA/CA
I know there exists another type of CSMA/CA that is entirely deterministic, ie that cannot lead to collision even if both ends speak at exactly the same time. It is used in some field networks, I cant remember which ones. I dont know enough of it, nor am i able to cite sources, so I describe it here hoping somebody will modify the article for me. thanks in advance to the one who will.
The principle is roughly: anybody wanting to speak first broadcasts its own address with '0' bits being recessive (much like tri-stated bits) on the medium; and re-reads its own address. If it reads something else than its address, it immediately stops and retries after the end of packet. This means the equipment with "11111.." address is able to speak whenever it wants to, and others follow, the highest address having the highest priority.
- - Ppchailley 16:59, 4 July 2006 (UTC)
- The sentence "all of the nodes on the bus are assigned an identification number or priority code." isn't completely correct
- in the context of CAN which is given as example for CSMA/BA. CAN doesn't assign identification numbers to nodes, but to
- message types. Therefor you can't prioritize an particular node but particular messages which could be send by any node.
- On the other some vendors use parts of the message ids as address informations. Unsignedvoid (talk) 08:00, 17 April 2008 (UTC)
Please check that you haven't got collision avoidance and collision detection mixed up. Some companies are definitely using "collision avoidance" to describe the random wait variants, which is logical - it doesn't detect a collision, it just avoids one being set up. So collision detection goes for the previous variant. —Preceding unsigned comment added by Neve Dan (talk • contribs) 18:30, 17 January 2008 (UTC)
- Ppchailley and "22.214.171.124" are correct -- he's describing the system used by Controller Area Network (and well described in that article). In that system, two simultaneous transmitters never garble a packet -- all receivers will hear a valid packet from either one or the other transmitter. --126.96.36.199 (talk) 05:08, 5 April 2008 (UTC)
- I added a few sentences. Does it make better sense now? --188.8.131.52 (talk) 05:08, 5 April 2008 (UTC)
The "on-board diagnostics" article mentions something called 'Carrier Sense Multiple Access with Non-Destructive Arbitration' (CSMA/NDA).
At first I thought it was talking about Controller Area Network. However, the article claims that CSMA/NDA was used on a "PWM" system long before CANbus.
If someone could find some reference to this "PWM" protocol (and no, the pulse-width modulation says nothing about automobile electronics protocols), that would be great. "CSMA/NDA" -- is it, or is it not, yet another synonym for CSMA/BA ? --184.108.40.206 (talk) 05:08, 5 April 2008 (UTC)
Does Bluetooth (802.15.1) really use CSMA-CA?
In "Protocols Using CSMA" it is stated that all 802.15 standards use the CSMA, but as far as I remember Bluetooth uses an master controlled channel access without CSMA. —Preceding unsigned comment added by 220.127.116.11 18.104.22.168 (talk) 07:34, 17 April 2008 (UTC)
The article states that in non-persistent CSMA the station "senses the medium continually until it becomes idle". This is a perfect description for 1-persistent CSMA, a very greedy type of CSMA. In non-persistent CSMA the node checks the medium at a moment T1 and if it finds it busy it checks again after a random time interval dT.
CSMA, CSMA/CD and CSMA/CA had been marked with a merge proposal since March 2008. No action had been taken. Material was repeated in these three article. The easy thing to do was to delete the repeated material in this article. I merge any useful edits into the CSMA/CA and CSMA/CA articles. I added Main tags from this article to CSMA/CA and CSMA/CD. This has worked out well. I have removed the merge tags. --Kvng (talk) 21:29, 27 April 2010 (UTC)
Harsh killa (talk · contribs) has added some information about virtual time CSMA. I am unable to verify the ref provided but other refs indicate that this is a technique that has been around since the mid-1980's. The best online description I found is here. --Kvng (talk) 11:42, 2 May 2012 (UTC)
Virtual time CSMA section is badly in need of clarification. I'm a CCIE with 20 years in the field and I had no idea what this meant when I read it. I will add it to my to-do list to provide clarity.Mtzweil (talk) 14:20, 2 September 2015 (UTC)
- Not seeing much of value, there. Hell, it ends with an ellipsis … ELLIOTTCABLE (talk) 07:39, 25 April 2017 (UTC)
For this page related to shared access media and related pages, I'd like for us to standardize on the following terms to help avoid confusion and ambiguity in describing the technology:
Use "Node" instead of "Station" Use "Transmit" instead of "Speak, Talk, of Send"