Le 15/12/2021 à 20:09, Carsten Bormann a écrit :
RFC 8930
It says:
VRB overcomes the limitations listed in Section 4. Nodes do not wait
for the last fragment before forwarding,
I did not to fragment, but to re-group.
Fragmentation and reassembly happening at the sender and receiver
respectively aims at dividing a big IP packet, such that to accomodate a
particular MTU.
Grouping and de-grouping would happen too at the sender and receiver
respectively, but its aim would be to create large packets, not small.
This group-degroup could help to reduce overhead on the network, and
level a burst of packets at extremely high bandwidth into a more gentle
sequenced flow.
An application generating very small packets sent over a huge bandwidth
Ethernet makes such problem.
Alex
_______________________________________________
Int-area mailing list
Int-area@ietf.org
https://www.ietf.org/mailman/listinfo/int-area
