Rules for delaying an ack:
* - delay time <= 0.5 HZ
* -we don't have a window update to send
* - must send at least every 2 full sized packets
* - must send an ACK if we have any out of order data
Now in this case if we have a 1MB file transmission on a local LAN then we
will only get good efficiency if we don't have to send acks every 2 full
sized packets.
In case of errors:
if it is a checksum error the packet is discarded and when the next packet
arrives it will be out of order and we will send an ack(along with a sack)
so the feedback will be delayed at most 1 packet time.
if the packet is lost then anyways the next packet will be ofo and we will
ack.
now the worst case can occur when we send two full sized packets and the
second packet is in error - then the ack will be received after the
delayed ack timer goes off - which is not too bad considering the delayed
ack timer goes off every 0.5Hz.
let me know what you think about this...
On Wed, 8 Dec 1999, David S. Miller wrote:
>
> If you delay the acks more than 2 packets you will see packet
> loss less quickly. Acks are the only feedback mechanism a senders can
> use to quickly respond to loss conditions. 2 is a compromise between
> wasting net bandwidth, taking advantage of the possibility that the
> ACK could piggyback with real data, and having reasonable reaction
> time to loss events.
>
> Later,
> David S. Miller
> [EMAIL PROTECTED]
>
Sandeep Agarwal
Electrical & Computer Engineering DearBorn Hall 208
Corvallis Phone: 541 737 2349
Fax: 541 737 1300 Mail:[EMAIL PROTECTED]
http://www.ece.orst.edu/~sandeep
-
To unsubscribe from this list: send the line "unsubscribe linux-net" in
the body of a message to [EMAIL PROTECTED]
