IFS

IFS - Inter-Frame Spacing is a parameter used in Ethernet, and it represents the minimum amount of time between two transmissions. If there is no IFS between the transmissions, it is likely to cause a collision error - which might be a cause of corrupted data.

In this flow, the IFS is used after the decision of whether the channel is free and the CTS(Clear to send) is received before the time-out of the timer. 

But then if the transmission starts right after the previous transmission, why should it be a issue of collision where there is no overlap between the two?

This is because the overlap is not the only factor of the collision between transmissions. Other factors such as network congestion, bandwidth limitations, and the processing power of the receiving device can also affect the ability of the network to handle multiple transmissions simultaneously. Therefore, it's important to include a certain amount of delay between transmissions to ensure that the network can handle them efficiently and reliably. This delay is typically implemented using techniques such as backoff algorithms or randomized delays.

Next are some examples of backoff algorithms.

1. Binary Exponential Backoff: In this algorithm, the sender waits for a random amount of time before retransmitting the packet. If there is a collision again, the sender doubles the waiting time and retries. This process is repeated until the maximum retry limit is reached.
2. Truncated Binary Exponential Backoff: This algorithm is similar to binary exponential backoff, but with a maximum waiting time limit. Once the waiting time reaches the maximum limit, the sender stops retrying and reports an error.
3. Adaptive Exponential Backoff: In this algorithm, the waiting time between transmissions is determined based on the number of collisions that have occurred. The waiting time increases exponentially with the number of collisions.
4. Linear Backoff: In this algorithm, the waiting time between transmissions is increased linearly after each collision. This means that the sender waits for a fixed amount of time after the first collision, and then increases the waiting time by a fixed amount after each subsequent collision.
5. Full Jitter Backoff: In this algorithm, the sender selects a random number between 0 and the maximum waiting time, and waits for that amount of time before retrying. This randomization helps to avoid synchronization between competing senders.