-----Original Message-----Hi Jonathan
From: Baktha Muralidharan [mailto:muralidb@cisco.com]
Sent: Thursday, June 13, 2002 12:51 PM
To: Jonathan Lang
Cc: ccamp@ops.ietf.org
Subject: Comments on draft-ietf-ccamp-lmp-04.txt
Please see my comments on ccamp-lmp-04 below, prefixed Baktha>
Regards,
/Baktha
==========================================================================
1.
"...
10.2.
Retransmission Algorithm
After a node transmits a message requiring acknowledgement, it
should immediately schedule a retransmission after Ri seconds. If a
corresponding acknowledgement message is received before Ri seconds,
then message retransmission SHOULD be canceled. Otherwise, it will
retransmit the message after (1+Delta)*Ri seconds. The
..."
Baktha> Shouldn't the first retransmission be after Ri seconds, rather than (1+Delta)*Ri ?
2."...
Prior to initial transmission, initialize Rk = Ri and Rn = 0.
while (Rn++ < Rl) {
transmit the message;
wake up after Rk milliseconds;
Rk = Rk * (1 + Delta);
}
/* acknowledged message or no reply from receiver and Rl
reached*/..."
Baktha> It is not clear if the initial transmission happens before entering the [while] loop or
Baktha> inside the loop. If the initial transmission happens happens inside the loop, we want
Baktha> the while condition to be (Rn++ <= Rl)
3.
Baktha> I assume that the backoff procedure applicable to ALL LMP procedures; might want to state this explicitly.
4.
Baktha> If the response to an LMP message is received within "rk" milliseconds, it is accepted and the retransmission is cancelled.
Baktha> I believe that this renders the timeout period insignificant.
5.
Baktha> State that the retry limit is a local configuration parameter.