An idea: I just happened to be browsing the 802.11b IEEE documentation today.
802.11b does retransmission on layer 2. If the frame doesn't get delivered the first time, it attempts to resend it. I am not sure if there is a timeout mechanism. At 1500 bytes, packets are either not being successfully sent by the requesting host or they are not being transmitted successfully by the responding host. At 1200 bytes, some of the ping requests work, while some don't. Those that don't work are accumulated at layer 2 for retransmission. If the software creating the pings creates ping requests faster than layer 2 can service them, delay is added and pings will time out. You might be able to fix your problems by lowering the maximum frame size for layer2. This would force layer 2 to fragment the packets and then reassemble them. You could approximate this size by pinging probably. I assume there are other methods, but I am not familiar with them. ++JB On Tue, 16 Sep 2003 13:08:39 -0700, dtowers wrote > Hello all, > > I have a client on my WISP that this is occurring to. > > pinging a server at the center with various sizes > when pinging with packet size of 1500 I get request timed out (rto) > same rto down to around a packet size of 1200 bytes > at around 1200 byte packet there are about 5 successful pings then > goes into rto if I decrease packet size it will begin to be > successfull again for a while.. then it goes into rto again at which > point you must decrease packet size again and it begins to succeed agian > and on goes this pattern > > so packet size determines number of successful pings > the larger the packet the fewer successes > > Has anyone experienced this or have an idea what its cause would be? > > Derek Towers > -- > general wireless list, a bawug thing <http://www.bawug.org/> > [un]subscribe: http://lists.bawug.org/mailman/listinfo/wireless -JB -- general wireless list, a bawug thing <http://www.bawug.org/> [un]subscribe: http://lists.bawug.org/mailman/listinfo/wireless
