Indoor WLAN Design

Part III: Understanding the 802.11 MAC

Kjell Jrgen Hole
UiB
Last updated 17.01.07
Mail: [email protected]
URL: www.kjhole.com
Outline KJhole.com
What is Medium Access Control (MAC)?
Why you need a good understanding of the 802.11 MAC
Challenges for the MAC
MAC details
access technique
timing
frame format
Bridging
3.2
Denition of MAC KJhole.com
MAC Medium Access Control. The function in a wireless network
that arbitrates use of the network capacity and determines which
MSs are allowed to use the medium for transmission
3.3
Why You Need to Understand the MAC
A few years back most 802.11a,b,g networks had a relatively small
number of users and were rarely subjected to severe stresses. This
picture has changed. There now exist large networks with many
BSs and MSs
Tuning a wireless network is tied to parameters in the MAC spec-
ication. Knowledge of what those parameters do is needed to
understand the behavior of a network and the eects of tuning
RTS threshold
Fragmentation threshold
3.4
Understand the MAC KJhole.com
Device drives may expose low-level parameters. A good under-
standing of the MAC is needed to optimize these parameters
Troubleshooting a wireless network may be dicult. A packet
snier can be an invaluable aid. To take full advantage of the
snier, you need to understand what the packets mean to interpret
the networks behavior
3.5
Challenge for MAC: RF Link Quality
RF link quality Radio links in the Industrial Scientic and Medical
(ISM) band are subject to:
interference
noise
multipath fading
! To alleviate these problems, all transmitted frames must be ac-
knowledged
a packet is considered lost if no ack is received
3.6
Challenge for MAC: Hidden Node Problem
Nodes A and B both want to communicate with node C
Radio waves cannot reach full distance between A and B
From the perspective of node A, node B is a hidden node since
A cannot hear B
Simultaneous transmissions from A and B will collide at C
A B C
Area reachable
by node A
Area reachable
by node B
3.7
Hidden Node Problem KJhole.com
BSs and MSs are generally half-duplex transceivers, i.e., they
dont transmit and receive at the same time
dicult to detect collisions caused by hidden nodes
! As we shall see, the MAC utilizes Request to Send (RTS) and
Clear to send (CTS) signals to prevent frames from colliding
3.8
802.11 MAC Overview KJhole.com
All 802.11 (a,b,g) networks use the same MAC
The MAC layer sits on top of the physical layer
The MAC adapts Ethernet-style networking to radio links
The MAC uses Carrier Sense Multiple Access with Collision Avoid-
ance (CSMA/CA) to control access to the transmission medium
3.9
Distributed Coordination Function (DCF)
Ethernet-like CSMA/CA access is provided by the Distributed Co-
ordination Function (DCF)
DCF rst checks to see that the radio link is clear before trans-
mitting. To avoid collisions, MSs use a random backo after
each frame, with the rst transmitter seizing the channel
In some cases DCF uses RTS and CTS signals to further reduce
the possibility of collisions caused by hidden nodes
Remark: enhanced distribution functions (802.11e) add quality
of service functionality and introduce trac classes
3.10
RTS/CTS (1) KJhole.com
Assume that node A wants to send a frame to node C. Node A
initiates the process by sending an RTS frame. The RTS frame
has the following purposes:
it reserves the radio link
it silences any MS that hears it
If node C receives the RTS, it responds with a CTS. The CTS
also silences nodes in the vicinity
3.11
RTS/CTS (2) KJhole.com
When the RTS/CTS exchange is complete, node A transmits a
frame (containing user data) to node C
Node C must acknowledge the frame using a separate ACK frame
RTS
CTS
Frame
ACK
A C
3.12
Shortcoming of RTS-CTS Solution
time
A B C D
RTS
CTS CTS RTS
collision
packet
transmission
new
RTS
CTS
CTS
3.13
time
A B C D
RTS
CTS
CTS RTS
packet
transmission
from A to B
CTS
C did not hear B's CTS
since it was transmitting
its own RTS to D
packet
transmission
from C to D
Collision
Another illustration of the RTS-CTS problem
3.14
Shortcoming: Exposed Node Problem
Node C is transmitting to D
B overhears this, and is blocked
B wants to transmit to A, but is blocked by C
A B C D
Blocked node
3.15
Exposed Node Problem KJhole.com
To alleviate the exposed node problem, a node must wait a ran-
dom backo time between two consecutive new packet transmis-
sions time
3.16
RTS/CTS Overhead KJhole.com
Since the multiframe RTS/CTS transmission procedure consumes
capacity, it is only used for frames larger than the RTS threshold
Some device drivers allow you to control the RTS/CTS procedure
by setting the RTS threshold
3.17
Carrier-Sensing Functions KJhole.com
Carrier sensing is used to determine if the medium is available. Two
types of carrier-sensing functions exist:
Physical carrier-sensing Provided by the physical layer. Cannot pro-
vide all necessary information due to hidden nodes
Virtual carrier-sensing Provided by the Network Allocation Vector
(NAV). NAV is a timer used to indicate how long the medium
will be reserved. The NAV is sent from a transmitting station
to a receiving station. Other MSs count down from the NAV to
zero. When the NAV is nonzero, no frame is sent. A station also
updates its own NAV when it transmits a frame
3.18
NAV KJhole.com
Since an RTS frame is not necessarily heard by all MSs in the
vicinity, the receiving station sends a CTS frame with another
NAV
By using NAV, MSs can ensure that atomic operations, such as
an RTS/CTS sequence, are not interrupted
Sender
Receiver
NAV
RTS
CTS
Frame
ACK
NAV (RTS)
NAV (CTS)
time
3.19
Interframe Spacing KJhole.com
Three dierent interframe spaces are used to coordinate access
to the transmission medium
similar to Ethernet
create dierent priority levels for dierent types of frames
independent of transmission speed
3.20
Dierent Interframe Spaces KJhole.com
Short interframe space (SIFS) Used for the highest-priority trans-
missions, such as RTS/CTS frames and positive ack. High-
priority transmissions can begin once the SIFS has elapsed
DCF interframe space (DIFS) Minimum medium idle time for
contention-based services. MSs may have immediate access to
the medium if it has been free for a period longer than the DIFS
Extended interframe space (EIFS) Used only when there is an er-
ror in frame transmission
3.21
Contention-Based Access Using the DCF
Two basic rules apply to all transmissions using DCF:
I. A station with a frame to transmit senses the medium. Carrier
sensing is performed using both a physical medium-dependent
method and the virtual (NAV) method. If the medium is idle, it
waits to see if the medium remains idle for a time equal to DIFS.
If so, the station may transmit immediately.
II. If the medium is busy, the station waits for the channel to be-
come idle for the DIFS and prepares for an exponential backo
procedure (to be explained)
3.22
Additional Rules (1) KJhole.com
1. If the previous frame contained errors, the medium must be free
for the amount of the EIFS
2. Error recovery is the responsibility of the MS sending a frame.
Senders expect acknowledgments for each transmitted frame and
are responsible for retrying the transmission until it is successful
(a) positive acknowledgements are the only indication of success
(b) atomic exchanges must complete in their entirety to be suc-
cessful
(c) all unicast data must be acknowledged (not broadcast data)
3.23
Additional Rules (2) KJhole.com
3. When a station receives a medium reservation that is longer than
the current NAV, it updates the NAV. Setting the NAV is done on
a frame-by-frame basis
4. Packets larger than the RTS threshold must have RTS/CTS ex-
change and packets larger than the fragmentation threshold must
be fragmented
5. The following types of frames can be transmitted after the SIFS
and thus receive maximum priority: acknowledgments, CTS, and
fragments in fragmented sequences
3.24
Backo with the DCF (1) KJhole.com
After frame transmission has completed, and the DIFS has elapsed,
MSs may attempt to transmit congestion-based data. A period
called the contention window or backo window follows the
DIFS
Previous frame DIFS
. . .
Window slots
time
The backo window is divided into slots. Slot length is medium-
dependent; higher-speed physical layers use shorter slot times
3.25
Backo With the DCF (2) KJhole.com
Stations pick a random slot and wait for that slot before attempt-
ing to access the medium; all slots are equally likely selections.
The station that picks the rst slot wins
Contention window sizes are always 1 less than a power of 2 (e.g.
31, 63, 127, 255)
Each time the retry counter increases, the contention window
moves to the next power of two
3.26
Backo With the DCF (3) KJhole.com
The size of the contention window is limited by the physical layer
When the contention window reaches its maximum size, it remains
there until it can be reset
The window size is reset to its minimum size after a successful
frame transmission
3.27
Fragmentation and Reassembly KJhole.com
Higher-level packets and some large management frames must be
broken up into smaller pieces to t through the wireless channel
Fragmentation improves reliability in the presence of interference
A packet (or frame) is fragmented when the length exceeds the
fragmentation threshold
NAV is used to ensure that other MSs do not use the channel
during a fragmentation burst
3.28
Format of Generic MAC Frame KJhole.com
There exist three major types of MAC frames. We only consider
the basic frame structure in this lecture
The frame contains no less than four address elds. Not all frames
use all the addresses
Frame
control
Duration/
ID
Address 1 Address 2 Address 3
Seq-
ctl
Address 4 Frame body FCS
3.29
Frame Control Field KJhole.com
Frame control two-byte eld with 11 subelds:
Protocol version consists of two bits that indicate which version
of the 802.11 MAC is used. At present, only one version is
dened
Type and subtype elds identify the type of frame used. Ex-
amples are Beacon, RTS, and CTS frames
ToDS and FromDS bits indicate whether a frame is destined for
the Distribution System (DS), i.e., the wired backbone network
and the bridging functions in the BSs
3.30
More Frame Control Subelds KJhole.com
More fragment bit is set to 1 in all fragments, except the last,
when a higher-level packet is fragmented by the MAC
Retry bit is set to 1 when a frame is retransmitted
Power management bit indicates whether the sender will be in
power-saving mode after a frame is transmitted. A BS can never
be in power-saving mode
3.31
Even More Frame Control Subelds
More data bit is set by a BS to indicate to an MS in power-saving
mode that at least one frame addressed to the MS is buered at
the BS
WEP bit is set when data is encrypted with WEP
Order bit indicates whether fragments are transmitted in strict
ordering
3.32
Duration/ID Field KJhole.com
Duration/ID eld can be used to

set the NAV in microseconds, or

used by slumbering MSs to retrieve frames buered at the BS

Assuming that DCF is utilized

3.33
Address Fields

KJhole.com
Destination address is a 48-bit IEEE MAC identier corresponding
to the nal recipient of the frame
Source address is a 48-bit IEEE MAC identier that identies the
source of the transmission

There exist two more address elds not discussed here

3.34
Three Last Fields KJhole.com
Sequence control eld is used for both defragmentation and dis-
carding duplicate frames
Frame body contains the higher-layer payload
Frame check sequence is a Cyclic Redundancy Check (CRC)
3.35
Bridging (1) KJhole.com
A BS can be viewed as a bridge that translates frames between a
wired network and a wireless network
802.11-to-Ethernet bridge
When a BS receives a frame on the wireless interface, it checks for
basic integrity, removes a duplicate frame, decrypts content, re-
assembles fragments, and translates MAC header into a Ethernet
MAC header
the new frame is transmitted on the Ethernet interface
3.36
Bridging (2) KJhole.com
Bridging frames from the wired network to the wireless network
is quite similar to the reverse of the process just described
3.37
Summary KJhole.com
The 802.11 MAC is designed to provide Ethernet-style networking
to radio links
The DCF (Distributed Coordination Function) allows multiple in-
dependent MSs to interact without central control, and thus may
be used in an ad hoc network mode
It is likely that network administrators must have a good under-
standing of the MAC to troubleshoot heavy loaded networks
3.38