I keep wondering about a slight point about fragmentation and addresses fields...
If a station wishes to send large packets (better say MSDUs), it will fragment them for MPDU to match the threshold. OK. Say threshold is 300 bytes, then MSDU will be cut into 272 bytes pieces, plus 18 bytes 3 addresses fields and 10 bytes CRC and headers (control, duration and sequence).
272+18+10=300, everything is perfect.
But this station then has to know that the destination station will be on the wire, isn't it ? What I mean here, it that if AP has to relay this packet to another AP thru WDS, then it'll add a fourth address, thus exceeding threshold...
SO do station have to be aware of that, in their ARP requests for example ? Or do they always leave an extra 4 bytes in frag frames just in case ?
Hi Samarcande of France:
IEEE 802.11 fragmentation and reassembly takes place between a transmitter and its immediate receiver. If the immediate receiver is not the destination then a new fragmentation and reassembly may follow and may use a different fragmentation threshold and fragment size.
I hope this helps. Thanks. /criss