Cisco Networking: Network Traffic Basics
You know about traffic being sent out over a network but how exactly is this task performed. Here are the two main methods used to release traffic onto the network:
CSMA/CD: Carrier Sense, Multiple Access with Collision Detection
CSMA/CA: Carrier Sense, Multiple Access with Collision Avoidance
The impact of either of these methods is reduced greatly when working on switched networks without shared access devices, such as hubs or wireless access points (WAPs).
CSMA/CD is the main frame-control process on wired Ethernet networks. Here is a breakdown of the name:
Carrier sense is built into the network card. The network device can detect the carrier wave that is on the network. That goes along with the green light you get when the network is properly attached to the network.
Multiple access allows for multiple devices to share the networking medium through some process, communication standards, or protocol. Because only one node on the wire can really transmit data at a time, you need a sharing method that everyone or every host on the network agrees to.
Collision detection means that in addition to the multiple access sharing, you specifically use two factors to help determine whether a collision occurs:
Ethernet round trip time: The time it takes for a frame to reach every possible node in your collision domain.
Interframe gap: The minimum safe distance between frames on the network that allows time for the receive buffer (a small amount of memory on a network card used to store incoming data that needs to be processed) on the receiving stations to process all frames.
If a receiving station sees frames coming too quickly (that is, without a sufficient gap between them), it can send a jam signal to notify others of a collision. If a station sends a frame and sees more traffic before the round trip time has expired, it can send a jam signal as well.
When a jam signal or collision is sent on the network, all systems wait for a random period of time before attempting to send their data again. This random wait reduces the chance of another collision on a network.
To reduce the chance of collision from occurring at all, prior to sending data frames on the network, the system will listen on the network to see if there is anyone currently sending a frame. If there is current activity on the network, the system will wait for a break and then send its frames out on the network.
CSMA/CA is a similar system to CSMA/CD but instead of dealing with collisions when they happen, the goal is to have them not occur at all. Collision avoidance is the process used on the Macintosh AppleTalk network and is still used on 802.11 wireless networks.
So how does collision avoidance differ from collision detection? Well, it is that whole avoidance bit in the name. To avoid collisions:
Each station that wants to send data listens to the network to see whether the network is in use.
If the network is not in use, the station sends a Ready to Send (RTS) signal, hoping for a Clear to Send (CTS) signal back from the other stations.
After the other stations give the CTS signal, they are required to not send any data for a short length of time (milliseconds).
This leaves the network free for the sending station to use, which ensures that you have a near zero chance of having a collision when you send your data. It is actually more likely that a collision would occur when the RTS request is sent out on the network.