What are contention periods in CSMA? Why are they used?

Question

As far I know that, contention slots are required to avoid all channels from transmitting at the same time. But what happens when the contention slot is over? All stations will transmit together and it will lead to jamming of the channel then. What exactly is the role of contention slots and how does it avoid collisions?

Answer 1

Contention Period is basically the amount of time for which a Host should keep on transmitting data so that if there is a collision anywhere in the channel, the signal should reach the Source Host. Ideally Contention Period is taken as 2*(Propagation Delay) considering the worst case scenario that the collision happens at the destination node itslef. So in conclusion Transmission Time >= 2*Propagation Time.

As CSMA/CD donot have any Acknowledgement Mechanism it is difficult to identify any error related to Transmission. That is where the Contention Period is helpful for CSMA/CD to detect error in Transmission.

Comments

why is Tpd + 2*C*Tpd taken instead of 2Tpd in many cases? The efficiency value stands at 1/(1+6.44*Tpd) [taking C=e]

Answer 2

CSMA/CD

CSMA/CD stands for Carrier sense multiple access / Collision detection.

CSMA/CD is a technique used in sharing ethernet in which the sending computer has exclusive use of the cable during the transmission of the frame.

- All computers attached to Ethernet use CSMA/CD to co-ordinate their activities. 
- A computer wishing to transmit checks for electrical activity in the cable. This electrical activity is known as carrier.
- If there is no carrier computer transmits the frames. If there is it waits.
- However it is possible that two computers might simultaneously start the transmission. This interference is called collision.
- Sending computer monitors the signal on the cable and if it is different from the one it is transmitting it indicates that collision has occured and transmission stops immediately.
- Computer waits for the cable to become ideal before retransmission.
- For no further collision standard decides the maximum delay of d. Each computer chooses any random value less than d.
- In case collision occurs again delay becomes 2d, 4d and so on. This technique is called binary exponential backoff.