AD Evaluation comments on draft-ietf-rmt-bb-tfmcc-06

Hi,

I have done an AD evaluation of the draft and found a few things I like
to have clarified. Depending on the answer it might be necessary to
update the draft.

1. Section 1.3:
"TFMCC inherently works with all types of networks, including LANs, 
WANs, Intranets, the Internet, asymmetric networks, wireless networks, 
and satellite networks."

I reacted to the word "all" here. I think it needs a clarification that
it is for all networks that enables bi-directional traffic. Could this
be fixed in the text?

2. Section 1.3, last paragraph:

"TFMCC is designed for applications that use a fixed packet size, and
vary their sending rate in packets per second in response to congestion.
Some audio applications require a fixed interval of time between packets
and vary their packet size instead of their packet rate in response to
congestion.  The congestion control mechanism in this document cannot be
used by those applications."

Here my issue is with "Some audio applications require ...". I would
like to change this to "Some applications, e.g. those using audio, may
require ...". The problem is more encompassing than only audio, audio is 
however an good example.

3. Section 2.2.1, the ts_r in data packets. I think this section is
missing the addition of the delay between receiving the feedback packet
with the ts_r and the transmission of ts_r' in the sender to receiver path.

4. I also think the usage of ts_r and ts_i is make a bit unclear by not
separating the different instances of the timers. If I understand the
RTT measurements loops there are two one for the sender and one for each
receiver.

                   < ts_d >
ts_i             ts_now   ts_i2
                           ts_r'
    \             ^        \
     \           /          \
      \         /            \
      V        /             V
      t_now    ts_r          t_now2
               ts_i'
       < tr_d >


The sender loop as I understand it starts with the first message
containing ts_i. The receiver does a receiver timestamp on the arrival
t_now. Then when sending feedback it calculates ts_i'= ts_i + ts_r-t_now
This I think is understandable from the text. And following from that
the sender's measurement of RTT will be R_r = ts_now - ts_i'.

The receivers measurement loop is similar but uses ts_r which is the
local clock value of when the feedback messages is going to be sent.
According to issue 3 above there doesn't seem to be any indication that
ts_r' is ts_r + ts_d, where ts_d is ts_i2-ts_now. The receivers makes it
RTT calculation R_sample = t_now2 - ts_r'.

 From this I have the following questions:
A. Was it intentional to ignore ts_d? If not please clarify that the
sender should create a ts_r' when sending the value.

B. I think one should maybe clarify that there is a ts_now and tr_now as
they are different clocks. Now the same variable is used in both
chapters which is a bit confusing.

C. Clarify which value should be used in the calculations, the ts_r or
the ts_r'.

5. Section 3.2
" R_max = max(R_max, s/X + t_gran)"

Is it intentional to calculate s/X rather than 8s/X? I react to that the
value is bytes divided by bits/s and thus seem to contain a unit error.

6. Section 3.6: I am missing any definition on when slow start may be
used. What are the prerequisite for applying the slow start mechanism?

7. Section 4.3.3: In the D_r calculation it seems that the ts_r term
should be the ts_r' term instead. Secondly the R' calculation should not
contain the ts_i' value, rather ts_i value of the latest received data
packet. I think the used notation is creating confusion here.
Please clarify this.

8. Section 5.5: Could anyone please explain why the history discounting 
for I_tot1 does only include the DF_i and not also the DF value?

9. Split references to normative and informational.

 From my perspective the draft would benefit significantly from 
clarifying the RTT measurements variables and their usage. When doing 
that I think a the other more editorial things could also be fixed.

Cheers

Magnus Westerlund

Multimedia Technologies, Ericsson Research EAB/TVA/A
----------------------------------------------------------------------
Ericsson AB                | Phone +46 8 4048287
Torshamsgatan 23           | Fax   +46 8 7575550
S-164 80 Stockholm, Sweden | mailto: [email protected]