4.9.2.3 - WiFi, access control and collision avoidance

4.9.2.3 - WiFi, access control and collision avoidance

Learn how a device joins a wireless local network, what its security measures protect, and how shared-channel access reduces collisions. Follow both ordinary CSMA/CA and the optional RTS/CTS exchange.

Wireless adapters and access points

WiFi connects devices to a local network using radio communication rather than requiring a data cable to each device.

WiFi

A wireless local area networking technology based on international IEEE 802.11 standards.

A device needs a wireless network adapter, which encodes data for radio transmission and receives and decodes compatible radio signals. The adapter may be built into a laptop or phone or supplied as an external device. Both ends need compatible standards and settings.

In an infrastructure WLAN, a wireless access point provides the wireless connection for stations in its coverage area and bridges their traffic to the local network. It can connect wireless clients to wired services. Multiple access points can extend coverage. Walls, distance, interference and competing transmissions affect radio communication, so being in the building does not guarantee the same connection everywhere.

An access point's role differs from a router's role of forwarding between networks. A home “wireless router” often combines an access point, router and Ethernet switch in one box. The combination does not make WiFi synonymous with Internet access. Devices can use a WLAN to reach a local printer or file server even when the Internet connection is unavailable.

SSID and wireless security

A network name helps a user select a WLAN, while authentication and encryption determine who can join and what others can read.

The Service Set Identifier (SSID) is the identifier/name of a wireless network. An access point normally advertises it so that clients can find and select that network. Two access points may share an SSID to serve the same WLAN; an SSID is not the unique hardware address of an access point and is not a secret password.

MeasureIntended benefitImportant limitation
Strong transmitted-data encryption using WPA/WPA2Makes intercepted radio data unintelligible without the necessary keys; authentication controls participationRequires secure configuration and credentials; does not itself make an untrusted website safe
Disable SSID broadcastStops the network appearing in the usual list of advertised namesDoes not encrypt data or make the network undetectable; the identifier can be discovered from other wireless exchanges
MAC address allow listAllows only listed adapter addresses to associateMAC addresses can be observed and imitated/spoofed; an address is not proof of a person's identity

The specification names WPA and WPA2. WPA is a legacy scheme; WPA2 configured with strong encryption and suitable credentials provides stronger protection than relying on a hidden name or address list. These are different mechanisms, not three interchangeable forms of encryption. A weak shared passphrase can still undermine a protected WLAN.

For example, hiding a school's SSID may reduce casual attempts to select it, but a nearby observer can still detect radio activity. An allow list may prevent an unlisted device joining under its usual address, but cannot establish that an allowed address has not been copied. Encrypting transmitted data protects confidentiality against interception; properly managed authentication and keys protect access. A user must still consider endpoint security and application security.

CSMA/CA without RTS/CTS

Stations sharing a radio channel need rules that reduce the chance of transmitting over one another.

Carrier Sense Multiple Access with Collision Avoidance (CSMA/CA) allows many stations to use the medium while listening before transmission and delaying access to reduce collisions. “Avoidance” does not promise that collisions can never occur.

For a normal unicast exchange without RTS/CTS, the process is:

  1. The station senses/listens to the channel. If it is busy, it defers rather than deliberately transmitting over the existing exchange.
  2. Once the channel has been idle for the required interval, contention uses a randomly selected backoff measured in time slots. Stations that became ready together are less likely to transmit together because their countdowns can differ.
  3. The countdown progresses while the channel is idle. If another transmission makes it busy, the countdown pauses and resumes after the required idle interval.
  4. When the countdown reaches zero and access is permitted, the station transmits its data frame.
  5. The receiver sends an acknowledgement (ACK) for a successfully received unicast frame. If the sender does not receive the expected ACK, it retries using contention/backoff rules. A missing ACK may be caused by interference or loss; it is not proof of a particular collision.

Radio conditions make avoidance necessary: two stations can both judge the channel idle and begin at the same time, or a station may be unable to hear another transmitter that the receiver can hear. Random delays reduce repeated simultaneous attempts. A larger contention window on retries can spread attempts further apart, at the cost of additional waiting.

The ACK is sent as part of the controlled exchange after a shorter waiting interval; it is not a new ordinary data transmission competing in exactly the same way. The exact numerical intervals and radio modulation are outside this lesson's required depth.

RTS/CTS and hidden stations

RTS/CTS can reserve the shared medium before a longer data frame, including for stations that cannot hear the original sender.

Suppose laptops A and C can both communicate with access point B, but a wall prevents A and C hearing each other. They are hidden stations relative to one another. A senses no transmission from C, and C senses no transmission from A. Their simultaneous data frames may nevertheless interfere at B. Carrier sensing alone cannot solve this particular visibility problem.

The optional Request to Send/Clear to Send (RTS/CTS) exchange adds short control frames:

  1. After sensing and winning access through contention/backoff, A sends an RTS to B, announcing its requested exchange and duration.
  2. If B can grant the exchange, B replies with a CTS containing the remaining reservation duration.
  3. Stations hearing the reservation information defer for that duration. C may not hear A's RTS but can hear B's CTS, so it knows to wait while A sends to B.
  4. A sends its data; B acknowledges successful receipt. The reservation covers the controlled exchange rather than granting A permanent ownership of the WLAN.

The RTS and CTS are shorter than a long data frame. If contenders' requests interfere, less airtime is wasted than if both sent long data frames. But requests can still collide and other radio interference still exists. If an expected CTS or ACK does not arrive, the sender must retry under the access rules.

The extra control traffic takes time. RTS/CTS can therefore help with hidden stations or costly collisions on long frames while adding unnecessary overhead for some short exchanges. It is an optional addition to collision avoidance, not a replacement for listening or backoff.