AFP - An Adaptive Fragmentation Protocol Supporting Large Datagram Transmissions
Transferring datagrams is essential for a lot of tasks. If the data does not fit into one network packet, fragmentation is needed. The authors propose a fragmentation protocol that adapts to different MTUs and to the datagram size, ensuring efficient bandwidth utilization. The protocol is extensible to allow tailoring to network and application demands. Thus, it is suitable for deploying it in various scenarios and under different conditions, whereas other protocols do not scale well. They demonstrate the applicability on low-end embedded devices and in a tele-operated robot scenario. Additionally, they analyze the scalability of their protocol, evaluate the induced overhead for different MTUs and datagram sizes, and compare it to other fragmentation protocols.