Medium Access Control (MAC) Concepts and Architecture

The IEEE 802.11 MAC is common to all IEEE 802.11 PHY layers and specifies the functions and protocols required for control and access. The MAC layer is responsible for managing data transfer from higher-level functions to the physical media. Figure 2-4 illustrates this relationship to the Open Systems Interconnection (OSI) model.

Figure 2-4: IEEE 802.11 standards mapped to the OSI reference model

MAC Layer Services

Devices using the IEEE 802.11 PHY and MAC as part of a WLAN are called stations. Stations can be endpoints or APs. APs are stations that act as part of the DS and facilitate the distribution of data between endpoints. The MAC provides nine logical services: authentication, deauthentication, association, disassociation, reassociation, distribution, integration, privacy, and data delivery. An AP uses all nine services. An endpoint uses authentication, deauthentication, privacy, and data delivery. Each service utilizes a set of messages with information elements that are pertinent to the services. Table 2-2 describes these services.

Table 2-2: IEEE 802.11 MAC Services and Agents

MAC Service	Definition	Station Type
Authentication	Because WLANs have limited physical security to prevent unauthorized access, 802.11 defines the authentication services needed to control access to the WLAN. The goal of the authentication service is to provide access control equal to a wired LAN. The authentication service provides a mechanism for one station to identify another station. Without this proof of identity, the station is not allowed to use the WLAN for data delivery. All 802.11 stations, whether they are part of an IBSS or extended service set (ESS) network, must use the authentication service prior to communicating with another station.	Endpoint and AP
Open system authentication	This is the default authentication method, which is a very simple, two-step process. First, the station wanting to authenticate with another station sends an authentication management frame containing the sending station's identity. The receiving station then sends back a frame alerting whether it recognizes the identity of the authenticating station.
Shared key authentication	This type of authentication assumes that each station has received a secret shared key through a secure channel independent of the 802.11 network. Stations authenticate through shared knowledge of the secret key. Use of shared key authentication requires the implementation of encryption via the Wired Equivalent Privacy (WEP) algorithm.
Deauthentication	This type removes an existing authentication. The deauthentication service is used to eliminate a previously authorized user from any further use of the network. Once a station is deauthenticated, that station is no longer able to access the WLAN without performing the authentication function again. Deauthentication is a notification and cannot be refused. For example, when a station wants to be removed from a BSS, it can send a deauthentication management frame to the associated AP to notify the AP of the removal from the network. An AP can also deauthenticate a station by sending a deauthentication frame to the station.	Endpoint and AP
Association	Association maps a station to an AP and enables the AP to distribute data to and from the station. The association service is used to make a logical connection between a mobile station and an AP. Each station must become associated with an AP before it is allowed to send data through the AP onto the DS. The connection is necessary in order for the DS to know where and how to deliver data to the mobile station. The mobile station invokes the association service once and only once, typically when the station enters the BSS. Each station can associate with one AP, although an AP can associate with multiple stations.	AP
Disassociation	This breaks an existing association relationship. The disassociation service is used either to force a mobile station to eliminate an association with an AP or for a mobile station to inform an AP that it no longer requires the services of the DS. When a station becomes disassociated, it must begin a new association to communicate with an AP again. An AP may force a station or stations to disassociate because of resource restraints; the AP is shutting down or being removed from the network for a variety of reasons. When a mobile station is aware that it will no longer require the services of an AP, it may invoke the disassociation service to notify the AP that the logical connection to the services of the AP from this mobile station is no longer required. Stations should disassociate when they leave a network, although nothing in the architecture ensures this will happen. Disassociation is a notification and can be invoked by either associated party. Neither party can refuse the termination of the association.	AP
Reassociation	This type transfers an association between APs. Reassociation enables a station to change its current association with an AP. The reassociation service is similar to the association service, with the exception that it includes information about the AP with which a mobile station has been previously associated. A mobile station will use the reassociation service repeatedly as it moves throughout the ESS, loses contact with the AP with which it is associated, and needs to become associated with a new AP.
	By using the reassociation service, a mobile station provides information to the AP to which it will be associated and information pertaining to the AP to which it will be disassociated. This enables the newly associated AP to contact the previously associated AP to obtain frames that may be waiting there for delivery to the mobile station as well as other information that may be relevant to the new association. The mobile station always initiates reassociation.	AP
Privacy	This type prevents the unauthorized viewing of data through the use of the WEP algorithm. The privacy service of IEEE 802.11 is designed to provide an equivalent level of protection for data on the WLAN as that provided by a wired network with restricted physical access. This service protects that data only as it traverses the wireless medium. It is not designed to provide complete protection of data between applications running over a mixed network. With a wireless network, all stations and other devices can hear data traffic taking place within range on the network, seriously impacting the security level of a wireless link. IEEE 802.11 counters this problem by offering a privacy service option that raises the security of the 802.11 network to that of a wired network. The privacy service, applying to all data frames and some authentication management frames, is an encryption algorithm based on 802.11.	Endpoint and AP
Distribution	This authentication provides data transfer between stations through the DS. Distribution is the primary service used by an 802.11 station. A station uses the distribution service every time it sends MAC frames across the DS. The distribution service provides the distribution with only enough information to determine the proper destination BSS for the MAC frame. The three association services (association, reassociation, and disassociation) provide the necessary information for the distribution service to operate. Distribution within the DS does not necessarily involve any additional features outside of the association services, although a station must be associated with an AP for the distribution service to forward frames properly.	AP
Data delivery	This provides data transfer between stations.	Endpoint and AP
Integration	This provides data transfer between the DS of an IEEE 802.11 LAN and a non-IEEE 802.11 LAN. The station providing this function is called a portal. The integration service connects the 802.11 WLAN to other LANs, including one or more wired LANs or 802.11 walls. A portal performs the integration service. A portal is an abstract architectural concept that typically resides in an AP, although it could be part of a separate network component entirely. The integration service translates 802.11 frames to frames that can traverse another network.	AP
Source: Intelligraphics and LaRocca, 135

MAC Layer Architecture

As illustrated in Figure 2-4, both the PHY and MAC layers are conceptually divided into management and data transfer capabilities. The PHY management capability is provided by the PHY layer management entity (PLME). The MAC management capability is provided by the MAC layer management entity (MLME). The PLME and the MLME exchange information about PHY medium capabilities through a Management Information Base (MIB) (see the following paragraphs for more information). This is a database of physical characteristics such as possible transmission rates, power levels, and antenna types. Some of these characteristics are static and some can be changed by a management entity. These management functions support the main purpose of the MAC, which is to transfer data elements. These data elements originate in the Logical Link Control (LLC) layer. Packages of data passed to the MAC from the LLC are called MAC service data units (Medusa). In order to transfer the Medusa to the PHY, the MAC uses messages (frames) containing functionality-related fields. Three types of MAC frames are available: control, management, and data. One of these messages is called a MAC protocol data unit (MPDU). The MAC passes MSDU to the PHY layer through the Physical Layer Convergence Protocol (PLCP). The PLCP is responsible for translating Medusa into a format that is physical medium dependent (PMD). The PMD layer transfers the data onto the medium.

MAC data transfer is controlled through two distinct coordination functions. The first is the distributed coordination function (DCF), which defines how users contend for the medium as peers. DCF data transfers are not time sensitive and delivery is asynchronous. The second is the point coordination function (PCF), which provides centralized traffic management for data transfers that are sensitive to delay and require contention-free access.^[10]

Management Information Base (MIB) 802.11 contains extensive management functions to make the wireless connection appear much like a regular wired connection. The complexity of the additional management functions results in a complex management entity with dozens of variables. For ease of use, the variables have been organized into an MIB so that network managers can benefit from taking a structured view of the 802.11 parameters. The formal specification of the 802.11 MIB is Annex D of the 802.11 specification. The 802.11 MIB is designed by the 802.11 Working Group.^[11]

Distributed Coordination Function (DCF) The DCF defines how the medium is shared among members of the wireless network. It provides mechanisms for negotiating access to the wireless medium as well as mechanisms for reliable data delivery. One of the fundamental differences between wired and wireless media is that it is difficult to detect and manage data collisions on wireless media. The primary reason for this is that stations in a radio network are not guaranteed to hear every other station's transmissions. This is typically the case when an AP is used in IEEE 802.11's infrastructure BSS and is called the hidden-node problem.

Point Coordination Function (PCF) The PCF polls associated stations and manages frame transmissions on their behalf. A station performing PCF traffic management is called a point coordinator (PC). The PCF is an optional capability that provides connection-oriented services for delay-sensitive traffic. The PCF is more complex to implement, but it provides a moderate level of priority frame delivery for time-sensitive transmissions.

The PC uses beacon signals to broadcast for the duration of a contention-free period to all associated stations. This causes them to update their network allocation vector (NAV) and wait for the duration of the contention-free period. In addition, stations must wait for the PCF interframe space (PIFS) interval to further decrease the possibility of data collisions. The transmission of the additional polling and ACK messages required by the PCF is optimized through piggybacking multiple messages in a single transmission. For example, the PC may append both ACKs of previous transmissions and polling messages for new traffic to a data frame. This enables the transmission to avoid waiting for the interframe interval specified for individual frame transmissions.^[12]

The basic access method for 802.11 is the DCF, which uses CSMA/CA. This requires each station to listen for other users. If the channel is idle, the station may transmit. If the station is busy, it waits until transmission stops and then enters into a random backoff procedure. This prevents multiple stations from seizing the medium immediately after completing the preceding transmission.

Packet reception in DCF requires acknowledgement, as shown in Figure 2-5. The period between the completion of packet transmission and the start of the ACK frame is one short interframe space (SIFS). ACK frames have a higher priority than other traffic. Fast acknowledgement is one of the salient features of the 802.11 standard, because it requires ACKs to be handled at the MAC sublayer.

Figure 2-5: CSMA/CA backoff algorithm

Transmissions other than ACKs must wait at least one DCF inter-frame space (DIFS) before transmitting data. If a transmitter senses a busy medium, it determines a random back-off period by setting an internal timer to an integer number of slot times. Upon expiration of the DIFS, the timer begins to decrement. If the time reaches zero, the station may begin transmission. If the channel is seized by another station before the timer reaches zero, the timer setting is retained at the decremented value for subsequent transmission. This method relies on the physical carrier sense. The underlying assumption is that every station can hear all the other stations.^[13]

^[10]LaRocca, 134-135.

^[11]Gast, 383.

^[12]LaRocca, 140-141.

^[13]Zyren and Petrick.