ELEC3506

Topics

Wireless NetworksLecture 11 PDF15 min

802.11 MAC and CSMA/CA

The two 802.11 MAC sublayers, why DCF uses collision avoidance instead of detection, the RTS, CTS, data, ACK exchange with its inter frame spaces, the NAV, and how PCF polls on top.

By the end of this page you should be able to

  • Name the two 802.11 MAC sublayers and say which one uses CSMA/CA
  • Give the three reasons CSMA/CD cannot be used in a wireless system
  • Write out the CSMA/CA sequence from channel sensing to ACK, with the wait before each frame
  • Explain how the NAV avoids collisions and what happens when RTS frames collide
  • Say why PCF can only run in an infrastructure network and how it gets priority over DCF

The idea

A wired station can listen while it sends and notice a collision. Radio gives it nothing to notice, so 802.11 does the opposite. Stations announce what they are about to do and everyone else keeps quiet for the announced time. A collision becomes something to avoid, not something to find.

CSMA/CD and the wired case are taught in Module 4, and the first look at CSMA/CA and RTS/CTS is there too. This topic is Lecture 11’s version of CSMA/CA, with the timing sequence and the PCF alternative.

How it works

Two MAC sublayers

IEEE 802.11 defines two MAC sublayers.

  • Distributed coordination function (DCF) uses CSMA/CA.
  • Point coordination function (PCF) is a centralised, contention-free polling access method.

How it works

Why DCF cannot use CSMA/CD

The lecture gives three reasons.

  • In wireless systems, collisions do not add enough energy for stations to detect them effectively.
  • The hidden node problem. A station may not hear the sender it collides with.
  • Signal fading could prevent a station at one end from hearing a collision at the other end.

So collisions are avoided with the CSMA/CA strategy.

How it works

The CSMA/CA sequence

  1. The source senses the channel until it is idle.
  2. After finding an idle channel, the source waits a period called the distributed inter frame space (DIFS).
  3. The source sends a control frame called request to send (RTS) to the destination.
  4. After receiving the RTS, the destination waits a short inter frame space (SIFS), then sends a control frame called clear to send (CTS) to the source.
  5. The source waits another SIFS and sends the data.
  6. The destination waits another SIFS and sends an ACK.

Only the first wait is a DIFS. Every wait after that is a SIFS. The slide that draws this timeline is a figure and extracted to nothing, so the sequence above is taken from the bullet list on the slide before it. The lecture gives no values in microseconds for DIFS, SIFS or the time slot on these slides. The 50, 20 and 10 microsecond figures that appear in Module 4 come from that module’s lecture.

How it works

Network allocation vector (NAV)

Collision avoidance works as follows. When a station sends an RTS, it includes the duration of time it needs to occupy the channel. Other affected stations create a timer called the network allocation vector (NAV). The NAV shows the time that must pass before those stations are allowed to check the channel for idleness. Each time a station accesses the system and sends an RTS, the other stations start their NAV.

Collision during the handshake. Two or more stations may send RTS frames at the same time and those frames may collide. The sender assumes there may have been a collision because it does not receive a CTS. It uses the back-off strategy and tries again.

How it works

Point coordination function (PCF)

PCF is a centralised, contention-free polling access method.

  • It is implemented over the DCF and used for time-sensitive transmissions.
  • It can only be implemented in an infrastructure network, not in an ad hoc network.
  • The AP performs the polling for the stations.
  • To give PCF priority over DCF, another set of inter frame spaces is defined, PIFS and SIFS. SIFS are the same as in DCF. PIFS (PCF IFS) are shorter than DIFS. If a station wants to use only DCF and the AP wants to use PCF at the same time, the AP gets priority.
DCFPCF
Access methodCSMA/CA, contention-basedCentralised polling, contention-free
Works in ad hoc networkYesNo
Who controls accessEvery stationThe AP
Wait before sendingDIFSPIFS, shorter than DIFS
Intended forNot stated on the slidesTime-sensitive transmissions
The two 802.11 MAC sublayers as the lecture describes them.

Where marks get lost

Silence is the only collision signal

The source never sees a collision. A missing CTS stands in for one, and the source backs off and tries again. Do not write that the source detects the collision, because the lecture’s first reason for CSMA/CA is that detection does not work.

In the exam

  • Name both sublayers and what they do. DCF with CSMA/CA, PCF with centralised polling.
  • Three reasons CSMA/CD cannot be used. Weak collision energy, hidden node, fading.
  • Sequence with its waits. Sense idle, DIFS, RTS, SIFS, CTS, SIFS, data, SIFS, ACK.
  • NAV. Set from the duration in the RTS. Other stations may not check the channel until it expires.
  • No CTS means a possible RTS collision. Back-off and retry.
  • PCF is infrastructure-only, the AP polls, and PIFS is shorter than DIFS.

Check yourself

  • DCF uses CSMA/CA and PCF is polled by the AP.
  • Wireless cannot detect collisions, so the exchange is RTS, CTS, data, ACK, each separated by a SIFS.
  • Other stations hold off for the NAV set from the RTS duration.
  • PCF only exists in infrastructure networks and wins against DCF because PIFS is shorter than DIFS.

Check yourself

  1. Which of these is NOT one of the lecture's reasons CSMA/CD cannot be used in a wireless system?
  2. In which order does the CSMA/CA exchange go once the channel is sensed idle?
  3. What does a station do on seeing an RTS that is not addressed to it?
  4. Why does an AP using PCF win against a station using only DCF at the same moment?
  5. Can PCF be used in an ad hoc network?