Talk:Carrier sense multiple access with collision avoidance

From Wikipedia, the free encyclopedia
Jump to: navigation, search
WikiProject Computing (Rated Start-class)
WikiProject icon This article is within the scope of WikiProject Computing, a collaborative effort to improve the coverage of computers, computing, and information technology on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.
Start-Class article Start  This article has been rated as Start-Class on the project's quality scale.
 ???  This article has not yet received a rating on the project's importance scale.

shamefully deleted my own comment. I was wrong. --Alvestrand 20:27, 18 January 2006 (UTC)

Exposed terminal[edit]

The hidden terminal problem was mentioned, but no word about the exposed terminal problem. —Preceding unsigned comment added by (talk) 10:18, 1 July 2008 (UTC)


Can a collission arise while sending a jamming signal?

As noted in the article, for radio systems there is often the possibility that the jamming signal or data transmission will not be detected by another radio, resulting in a collision. This is even more likely for mobile radio systems where often the radio path to another radio may be blocked by terrain effects.--Rjstott 06:59, 3 July 2006 (UTC)

CSMA/CA does not use RTS/CTS[edit]

The article incorrectly states that CSMA/CA uses RTS/CTS. That is an improvement in the MACA protocol and is optional in 802.11 protocol.

  • You are correct in stating that the IEEE 802.11 Standard does not specify as mandatory the use of the RTS/CTS control-fame handshaking MAC dialogue protocol mechanism. However, 802.11 is not the definitive form of CSMA/CA, which is actually TWO protocols in one.
    • The first is the carrier sense - CS' mechanism (as used by 802.3 Ethernet) by which the MAC layer process first listens to (senses) the physical layer (802.3 - wire, cable; 802.11 - local radio space) before transmitting.
      • An important difference between the two physical layers used by (802.3 Ethernet & 802.11 RF) follows.
      • 802.3 Ethernet. A device (A) that is transmitting to another device (B) is able to listen to (sense) the cable while transmitting and thus detect a collision caused by an interfering device (C) on the local subnet directly. That is, Device A is able to receive the signal that it is putting onto the wire as well as the signal from device C. As does device B, but the superimposed signals are unintelligible and need to be re-transmitted which is why collisions are a source of data link inefficiency. At least both device A and C are aware that they need to re-transmit their respective frames.
      • 802.11 RF. For the same situation above except that devices now have a radio frequency (RF) transceivers rather than cable. Because of the way that electromagnetic radiation propagates and the signal attenuates with distance. The radio signal that node A receives from interfering node C is many orders of magnitude (10's of dB's) less than the signal that itself (node A) is also transmitting. This means that node A is not able to discover that collision has occurred. This is why 802.11 employs acknowledgment fames (ACK) i.e. A node is thus able to detect a collision by the non-receipt of an ACK frame that would have been transmitted by node B if it had been able to make sense of the frame that node A had transmitted.
    • The second is the collision avoidance - CA mechanism that Phil Karn gave us in MACA. Because of the nature of the wireless medium, there are situations in which better link performance results by nodes utilizing a short control frame (RTS/CTS) handshake to explicitly alert all nearby potentially interfering nodes of the impending DATA frame tranasmission.
  • Therefore you are incorrect in stating that CSMA/CA does not use RTS/CTS, as this handshake is the CA part of CSMA/CA. DrBob127 (talk) 03:09, 22 November 2008 (UTC)

[NOTE] I propose that I incorporate the above distinction into the article itself. Any takers...??? DrBob127 (talk) 02:06, 6 December 2008 (UTC)

Jamming signals[edit]

It says in the "Usage" sub-heading: Apple's LocalTalk implemented CSMA/CA on an electrical bus using a three-byte jamming signal

What exactly is this jamming signal? Isnt jamming signal sent when a collision is detected in the network to warn the other hosts connected to the CSMA network about the collision? If so, doesnt this make it CSMA/CD rather than CSMA/CA? —Preceding unsigned comment added by (talk) 02:29, 22 September 2007 (UTC)

I found the answer to these questions here: Please feel free to delete this post if you are an authorized user (I am not really familiar with the system). —Preceding unsigned comment added by (talk) 10:47, 25 September 2007 (UTC)

  • Answer to your question: collisions can occur in CSMA/CA when two nodes start sending at the same time.

Station Checking[edit]

It says the station "checks to see if the channel is still free". Does anyone know exactly how it does this? I can't find anywhere how the station actually checks whether the line is busy. A voltage signal would run into the same collision problems. Anyone know? —Preceding unsigned comment added by (talk) 02:18, 13 June 2008 (UTC) 02:40, 13 June 2008 (UTC)

I'd like to know this as well. Could the client possible just send the signal that the station senses ahead of the packet? –Sigmarz talkedits 10:27, 23 June 2008 (UTC)

The MAC layer has access to a Receive Signal Strangth Indicator (RSSI) which is used to give the MAC visibility of the current status of the PHY. DrBob127 (talk) 03:44, 22 November 2008 (UTC)

Figure modification and performance section deletion[edit]

After reading the article I feel that it needs substantial improvement. The figure is misleading, since the transmit application data box is in the RTS/CTS side, thus implying that no application data is transmitted unless RTS/CTS is used.

I would remove the performance section altogether. DSSS and FHSS are phy layer techniques and they are somewhat orthogonal to the layer-2 protocol. In fact, new devices use ofdm or mimo-ofdm at the phy layer. The widely cited paper of Bianchi in jsac 2000 offers a performance analysis of dcf, which is a particular implementation of csma/ca. Jbarcelo (talk) 10:42, 31 October 2010 (UTC)

Optional RTS/CTS - Concept vs. standard confusion![edit]

I believe that sometimes CSMA/CA refers to the concept itself (adding RTS/CTS to CSMA) and sometimes it refers specifically to the CSMA/CA defined in the 802.11 family (not sure when or in which ammendment). However the RTS/CTS part is optional in the 802.11 standard. I am not sure if it means that the manufacturer does not even need to implement it (don't think so) or if it only means that you can disable it in the settings for your 802.11 networking card or AP? So when people read that CA is optional in CSMA/CA, it refers to the choice by IEEE to make the implementation of RTS/CTS optional.

As always Tanenbaum has an explanation that makes sense. From 'Computer Networks', 5th edition, section 4.4, pg. 324.

However, while RTS/CTS sounds good in theory, it is one of those designs that has proved to be of little value in practice. Several reasons why it is seldom used are known. It does not help for short frames (which are sennt in place of the RTS) or for the AP (which everyone can hear, by definition). For other situations, it only slows down operation. RTS/CTS in 802.11 is a little different than in the MACA protocol we saw in Sec 4.2 because everyone hearing the RTS or CTS remains quiet for the duration to allow the ACK to get through without collision. Because of this, it does not help with exposed terminals as MACA did, only with hidden terminals. Most often athere are few hidden terminals, and CSMA/CA already helps them by slowing down stations that transmit unseccessfully, whatever the cause, to make it more likely that transmission will succeed

However if talking about the concept of CSMA with CA, of course CA an inherent defining feature. (talk) 20:07, 11 May 2012 (UTC)

Not only wireless[edit]

It is important to concentrate on the protocol itself in the article. Yes, wireless is the most common current use, but that is by no means an intrinsic part of the protocol itself. Snori (talk) 18:40, 26 June 2012 (UTC)

CS versus CA[edit]

If Carrier sense multiple access (CSMA) " a probabilistic Media Access Control (MAC) protocol in which a node verifies the absence of other traffic before transmitting on a shared transmission medium", then what does /CA add? This CSMA/CA article currently states that "...carrier sensing is used, but nodes attempt to avoid collisions by transmitting only when the channel is sensed to be "idle"..." - but surely this is true of CSMA/CD too? Snori (talk) 18:57, 26 June 2012 (UTC)

Reason for using CSMA/CA over CSMA/CD[edit]

The article says that CSMA/CA is used in radio networks because CSMA/CD cannot solve the hidden terminal problem. While this is correct, the more important reason is imho that in radio networks, nodes are typically not able to transmit and receive at the same time (cf. also section "CSMA/CA does not use RTS/CTS"). Thus, it is typically not possible to implement CSMA/CD in a wireless network. This should imho be mentioned in the article. (talk) 13:52, 6 October 2013 (UTC)