[Refer to drawings on board that accompany these notes.]
The
first Ethernet, Ethernet DIX, was named
after the companies that proposed it:
Digital,
Intel, and Xerox. During this time, the
Institute of Electrical Engineers (IEEE) had been working on Ethernetr
standardization, among other things,
which became know as project 802.
Upon its success the Ethernet plan evolved into the IEEE 802.3
standard. Based on carrier sensing, as
originally developed by Robert Metcalfe, David Boggs and their team of
engineers, Ethernet became a major player in communications mediums, competing
head-to-head with IBM’s Token Ring technology or IEEE 802.5.
When a station on an Ethernet network is ready to transmit, it must first listen for transmissions on the channel. If another station is transmitting it is said to be “producing activity”. This activity or transmission, is often called a carrier. (See board about voltage explanation). This is how Ethernet became known as the “carrier sensing communications medium”. With multiple stations, all sensing carriers on an Ethernet network, this mechanism was called Carrier Sense with Multiple Access or CSMA.
If a carrier is detected by a station, the station (PC or server or whatever) will wait for at least 9.6 micro-seconds after the last frame passes before transmitting its own frame. When two stations transmit simultaneously, a “fused signal bombardment”, otherwise known as a collision occurs. Ethernet stations detect collisions to minimize problems. This technology---software algorithms---was added to CSMA to become Carrier Sense with Multiple Access and Collision Detection or CSMA/CD.
Stations that participate in the collision immediately abort their transmissions. The first station to detect a collision sends out a “all stations alert” that a collision is occurring. At this point ALL stations execute a random collision timer to force a delay before transmitting their frames. This timing delay is called (technically) the “back-off algorithm”. If multiple collisions are detected the random delay timer is doubled. Hence, a large, busy network on HUBS can have serious performance issues when multiple on-going collisions are present. There is a “point of no return” that if after 10 consecutive collisions and multiple double random delay times, performance will not improve significantly. Hackers, early-on, used this “Ethernet flooding” method to impact a network.
Review the following URL to help your understanding of CSMA/CD:
http://www.erg.abdn.ac.uk/users/gorry/course/lan-pages/csma-cd.html