We are having an issue with an application that send multicast streams with a packet length of 51KB. Obviously the packet get's fragmented in around 34 fragments of 1514 bytes. The first issue we saw was that the FWSM only allow up 24 fragments per IP packet, so we increased it to fit the need. Then we saw that the packets before going through the FWSM are 1514 bytes length. After the packets go through the FWSM( which will be assemble, inspected, fragmented and forward) the length of packets are around 1114 bytes, which means more fragmented packets. Is this a normal behavior of the FWSM or could it be a bug?
After doing more testing it seems that the FWSM can handle IP packets larger than 8KB. Anything over 8KB it's received out of order at client and because these is UDP, packets are drop by the client.