Looks like this is due to client SWS or Nagle override timeout in TCP.

Can you compare the packet headers for "with IM" and "without IM" case? The
Window and Acknowledgement fields are most interesting.

Max

All the flags, including ACK, and the Window values are the same, as between
the passthru-present and passthru-absent scenarios. The order of packets varies
slightly; for example, in regard to the first 3 packet exchanges,

In passthru-present:

1) Home3 -> Home2, S flag, window 0xFB04.
2) Home2 -> Home3, AS flags, window 0xFB04.
3) Home3 -> Home2, A flag, window 0xFB04.

In passthru-absent:

1) Home3 -> Home2, S flag, window 0xFB04.
2) Home3 -> Home2, A flag, window 0xFB04.
3) Home2 -> Home3, AS flags, window 0xFB04.

That is, the second and third packets are reversed in order, but the
directions, flags and window value are the same. Further, NetMon reports the 3
exchanges with the same timestamp, in passthru-present and in passthru-absent,
so maybe NetMon couldn't distinguish the order of the second and third packets
in each scenario.

"Maxim S. Shatskih" wrote:

--
If replying by e-mail, please remove "nospam." from the address.

James Antognini
Windows DDK MVP