What is the fragmentation threshold?

Answer 1

The maximum frame size wireless device can transmit without fragmenting the frame.

A collision happens when two devices that use the same medium transmit packets at exactly the same time. The two packets can corrupt each other, and the result is a group of unreadable pieces of data. If a packet results in a collision, the packet is discarded and it must be transmitted again. This adds to the overhead on the network and can reduce the throughput or speed of the network. Larger frames are more likely to collide with each other than smaller frames. You can make the wireless packets smaller by lowering the fragmentation threshold. If you lower the maximum frame size, it can reduce the number of retransmissions caused by collisions, and lower the overhead caused by retransmissions. However, making frames smaller introduces a different kind of overhead. Smaller frames introduce more overhead on the network too. This is especially true on a wireless network, because every fragmented frame sent from one wireless device to another wireless device requires the receiving device to acknowledge the frame. In times of high packet error rates (over five or ten percent collision or errors), lowering the fragmentation threshold can help improve performance of the wireless network. The time that is saved from reducing re-transmissions can be enough to offset the extra overhead added by using smaller packets. This can result in higher throughput. If the rate of packet error is low and you lower the fragmentation threshold, wireless network performance will decrease. This is because lowering the threshold adds protocol overhead and reduces protocol efficiency. The default maximum value 2346 is recommended.