Architecture and Functionality of a MAC Sublayer

Architecture and Functionality of a MAC Sublayer
Recall that the IEEE 802 family of standards has split the ISO/OSI data link layer into two parts:
The upper sublayer is the LLC sublayer, and the lower is the MAC sublayer (just above the PHY)
(as shown in Figure 4.1). This is in order to distinguish between medium access functionality and
other data link issues. Each IEEE 802 PHY standard (Ethernet, Token Ring, Token Bus, and so on)
specifies both the PHY aspects of the protocol as well as how medium access is to take place (as
shown in Table 4.6). For example, the IEEE 802.3 standard (Ethernet) specifies the media types that
can be used – a PHY issue – and specifies the use of the Carrier Sense Multiple Access/Collision
Detection (CSMA/CD) medium access protocol – a data link layer and MAC sublayer issue [453].
In contrast, the LLC sublayer manages to provide a single interface to the network layer for the numerous physical-layer topologies. This includes controlling the connection between sending and
receiving computers, and seeing that frames are transferred without errors [453].
One of the MAC services, the asynchronous data transfer service, manages the exchange of
data packets called MSDUs between devices (recall that every STA supports the MSDU delivery SS).
Technically,MSDUs themselves are not passed from device to device. The MSDU is the packet of data
going between the host computer’s software and the wireless LAN MAC [457]. An MSDU is typically
broken into smaller parts, each with a MAC header added, before encryption and transmission. This
process is known as fragmentation (discussed at the end of this section). These pieces of the original
MSDU are known as MAC Protocol Data Units (MPDUs). MPDUs are packets of data going between
the MAC and the antenna. For transmissions, MSDUs are sent by the operating system (OS) to the
MAC layer and are converted to MPDUs ready to be sent over the radio. For receptions, MPDUs
arrive via the antenna and are converted to MSDUs prior to being delivered to the OS [457]. If an
MPDU is lost in transmission, it can be resent instead of resending an entire MSDU.
All MAC frames share the same basic features: a MAC header for frame control, duration,
address, and sequence control information, a frame body (which varies by frame type), and a frame
check sequence (FCS) holding an IEEE 32-bit cyclic redundancy code (CRC). The FC field contains
protocol version, type, subtype, to DS, from DS, more fragments, retry, power management, more
data, WEP, and order subfields.
The 802.11 MAC supports CSMA/CA,2 implemented in all STAs, as its fundamental distributed
coordination function (DCF). This is almost the same DCF used in the IEEE 802.3 Ethernet
LANs – CSMA/CD (CSMA with collision detection). CSMA is a “listen-before-talk” protocol: STAs
“listen” to the transmission medium before sending a message. If the medium is in use, they use a
back-off algorithm to reschedule their transmission for a later time, when the medium could potentially
be free. Not all collisions are prevented with this scheme. If STA A sends a message, it will
take time (the propagation delay) before it reaches STA B. In the meantime, STA B may sense the
medium, not hear STA A’s message yet, deduce that the medium is free, and send a message that
collides with the first one. (On a LAN with an unusually long propagation period, or on a WAN,
the propagation time between stations may be too long for carrier sensing to do much good.) Additionally,
there is the “hidden terminal problem.” On a wireless network, STA C may be physically
prevented from ever hearing that STA A is transmitting, and constantly infer that it is safe to transmit
to STA B, initiating collision after collision. In a wired LAN, collisions are detected to make sure
messages involved in collisions are not lost for good, but time is lost and the medium is unnecessarily
tied up. Wired LANs can easily detect collisions by listening for voltage spikes on the transmission
medium. Wireless STAs cannot use this method because the signal of a transmitting STA dominates
over all other nearby signals, and competing signals may not be detected. One solution would be to  deploy expensive directional antennas and front-end amplifiers at each STA, with one set for transmitting
and one for receiving, to ensure that a STA could tell its transmitting antenna pattern from
its receiving antenna pattern. Arranging this situation is not convenient in radio terminals due to the
expense and the extra hardware encumbrance [454]. The collision avoidance (CA) method was developed
to serve alongside CSMA in wireless networks and is the basic access method adopted by the
802.11 standards. Under CSMA/CA, STAs monitor the transmission medium by both virtual and physical
means. The virtual monitor, the network allocation vector (NAV), is implemented in the MAC.
The NAV maintains a prediction of future traffic on the medium based on duration information that
is announced in RTS/CTS frames prior to the actual exchange of data. The duration information is
also available in the MAC headers of all frames sent during the CP other than the PS-Poll Control
frames. The physical monitor must be able to detect signals of certain types with certain degrees of
success [452].
Figure 4.6 provides an example of the operation of the CSMA/CA mechanism used in the IEEE
802.11 standard. Stations A, B, C, D, and E are engaged in contention for transmission of their
packet frames. Station A has a frame in the air when Stations B, C, and D sense the channel and
find it busy. Each of the three stations will run its random number generator to get a back-off time
at random. Station C followed by D and B draws the smallest number. All three terminals persist in
sensing the channel and defer their transmission until the transmission of the frame from terminal
A is completed. After completion, all three terminals wait for the interframe space (IFS) period and
start their counters immediately after completion of this period. As soon as the first terminal, Station
C in this example, finishes counting its waiting time, it starts transmission of its frame. The other
two terminals, B and D, sense C’s transmission and freeze their counters to the value that they have
reached at the start of transmission for terminal C. During transmission of the frame from Station C,
Station E senses the channel, runs its own random number generator that in this case ends up with a number larger than the remainder of D’s back-off time and smaller than the remainder of B’s, and
defers its transmission (until) after the completion of Station C’s frame. In the same manner as the
previous instance, all terminals wait for the IFS period and start their counters. Station D runs out of
its random waiting time earlier and transmits its own packet. Stations B and E sense D’s transmission,
freeze their counters and wait for the completion of the frame transmission from terminal D and the
IFS period after that before they start running down their counters. The counter for terminal E runs
down to zero earlier, and this terminal sends its frame while B freezes its counter. After the IFS
period following completion of the frame from Station E, the counter in Station B counts down to
zero and B sends its own frame [454].
The advantage of this back-off strategy over the exponential back-off used in IEEE 802.3 is that
the collision detection procedure is eliminated and the waiting time is fairly distributed in a way
that on average enforces a first-come, first-served policy [454]. Under CSMA/CA, priority levels can
be assigned to transmissions by shortening or lengthening the IFS time. Furthermore, all directed
traffic uses immediate positive acknowledgment (ACK frame) where retransmission is scheduled by
the sender if no ACK is received. CSMA/CA can be further refined by introducing RTS and CTS
frames after determining that the medium is idle and after any deferrals or backoffs under certain
circumstances (no RTSs are allowed prior to multicast and broadcast transmissions, for example,
because multiple, colliding CTSs would be the result).
After the DCF, the next important MAC function is also an access method: the point coordination
function (PCF). It is optional under the 802.11, but, if used, must coexist with the DCF. This access
method uses a point coordinator (PC), which should operate at the AP of the BSS, to determine
which STA currently has the right to transmit. The operation is essentially that of polling, with the
PC performing the role of the polling master. The PCF should distribute information within Beacon
management frames to gain control of the medium by setting the NAV in STAs. In addition, all frame
transmissions under the PCF may use an IFS that is smaller than the IFS for frames transmitted via
the DCF. The use of a smaller IFS implies that point-coordinated traffic should have priority access
to the medium over STAs in overlapping BSSs operating under the DCF access method.
The access priority provided by a PCF may be utilized to create a CF access method. The PC
controls the frame transmissions of the STAs so as to eliminate contention for a limited period of
time. When both the DCF and the PCF are in operation on the same BSS, they alternate, and CF
periods provided by the PCF alternate with contention periods under the DCF.
Another functionality of the MAC sublayer was brought up in the discussion of the asynchronous
data transfer service, one of the MAC services. This is the fragmentation of large MAC frames
into smaller MPDUs. Two types of MAC frames can be fragmented: MSDUs, which arrive in the
MAC via the LLC, and MAC management protocol data units (MMPDUs), which originate in the
MAC sublayer management entity (MLME) (discussed in the following text). When such frames are
about to be sent, they are fragmented if they are larger than a specified limit (stipulated in a variable
called a Fragmentation Threshold). Fragmentation creates MPDUs smaller than the original MSDU
or MMPDU length to increase reliability, by increasing the probability of successful transmission of
the MSDU or MMPDU in cases where channel characteristics limit reception reliability for longer
frames [452]. Also, since each individual MPDU is acknowledged, the smaller MPDUs are all that
needs to be resent if one is lost. Only unicast frames can be fragmented. The transmitting STA itself
carries out the fragmentation process, and each receiving STA reassembles the MSDUs/MMPDUs.
For historical reasons, the name for this recovery process is defragmentation.
The MAC data service translates LLC-style service messages into MAC-usable signals, and vice
versa. The 802.11 MAC sublayer is managed by a logical abstraction called the MLME (the PHY is
managed by an analogous unit called the PHY layer management entity (PLME)). The MLME provides
an interface for invoking MAC layer management functions. Furthermore, a station management entity
(SME) is resident in each STA. This is a layer-independent entity that may be viewed as residing in
a separate management plane or as residing “off to the side.” These entities interact with one another via logical service access points (SAPs). There is a
SME-MLME SAP, a SME-PLME SAP, and an MLME-PLME SAP. The 802.11 standards do not
specify the exact functions of SMEs. However, the services provided by the MLME SAP to SMEs are
defined. These include mundane information transfers revolving around requests by STAs to become
associated (disassociated, reassociated) and authenticated (unauthenticated) with the network, STA
power management and status, timer synchronization, and so on. Exchange of information concerning
physical matters (getting hardware devices to cooperate with each other) is facilitated via the PLME
SAPs [452].

802 standards and medium access protocols

Standard Medium access protocols
802.3 CSMA/CD
802.4 Token bus access
802.5 Token ring access
802.11 FHSS, DSSS, Infrared
802.11a OFDM
802.11b DSSS
802.11g OFDM