El 04/08/2015 a las 16:30, Black, David escribió:
Adrian,
Thanks for the response - this note contains the follow-ups on nits/editorial
items.
-----Original Message-----
From: Adrian Farrel [mailto:adrian(_at_)olddog(_dot_)co(_dot_)uk]
Sent: Monday, August 03, 2015 1:38 PM
Nits/editorial comments:
Section: 3.2.1 - Editorial suggestion: Changing "+" -> "+/-" in the
formula for nominal central frequency and re-defining n as a
(snip)
Ok, proof by (ITU-T) authority wins here.
Ramon> No change here then.
p.6 - please state that slot width is +/- wrt nominal central frequency.
Ah, took me a moment to see what you mean.
Yes, this could be clarified with
OLD
o Slot Width: The slot width determines the "amount" of optical
spectrum regardless of its actual "position" in the frequency
axis. A slot width is constrained to be m x SWG (that is, m x
12.5 GHz), where m is an integer greater than or equal to 1.
NEW
o Slot Width: The slot width determines the "amount" of optical
spectrum regardless of its actual "position" in the frequency
axis. A slot width is constrained to be m x SWG (that is, m x
12.5 GHz), where m is an integer greater than or equal to 1.
The slot width defines the amount of spectrum in use on
each side of the central frequency, thus the amount of
frequency in use is actually twice the value of the slot width.
That definitely helps.
Ramon> as noted by Fatai, I think the second part of the "NEW" is
incorrect. The slot width is always m*SWG. Do not confuse with NCF
granularity
I am also reluctant to add Fatai formula Frequency slot = [(central
frequency) - (slot width)/2] ~[(central frequency) + (slot width)/2]
since it is adding a new way of describing the slot in addition to the
(n,m) pair via the start / end NCF.
In short, IMHO the slot width is the amount of optical spectrum (m*SWG)
and the slot width is "centered" at its Nominal _Central_ Frequency (n).
That is it, regardless of whether SWG and NCF granularity happen to be
12.5 and 6.25
No changes here.
p.8 - Fig 4 could use a bit more explanation - the two frequency
slots occur at different points along the path.
Maybe...
OLD
o Effective Frequency Slot [G.870]: The effective frequency slot of
a media channel is that part of the frequency slots of the filters
along the media channel that is common to all of the filters'
frequency slots. Note that both the Frequency Slot and Effective
Frequency Slot are local terms.
NEW
o Effective Frequency Slot [G.870]: The effective frequency slot of
a media channel is that part of the frequency slots of the filters
along the media channel that is common to all of the filters'
frequency slots. Note that both the Frequency Slot and Effective
Frequency Slot are local terms.
Figure 4 shows the effect of combining two filters along a channel.
The combination of frequency slot 1 and frequency slot 2 applied to
the media channel is effective frequency slot shown.
END
That also helps.
Ramon> Changed to NEW. Also shifted the text "Frequency slot 1" of the
figure a bit to the left to be
centered
Nit: First nominal central frequency 'X' in Fig 5 needs to move 2
chars left.
I think it is one char :-)
Ramon> Changed, thanks.
Touche'
Section 4 - TE link term shows up w/o acronym expansion or definition.
Please define it before use.
Yes. Last line of section 4.
This section provides a mapping of the ITU-T G.872 architectural
aspects to GMPLS/Control plane terms, and considers the relationship
between the architectural concept/construct of media channel and its
control plane representations (e.g., as a TE link).
I don't understand how "e.g." defines "TE link".
Ramon> Changed to
NEW
(e.g., as a TE link, as defined in [RFC3945].)
Sections 4.2 and 4.3 - this may be my unfamiliarity, but it would have
helped to have some sort of heads-up at the start of the figures that
the top (non-GMPLS) portion of the figures prior to Figure 12 are
entirely in the optical domain. Perhaps explaining what the two
planes are (and how they're realized/implemented) in Figure 8 would help.
Hmmm. I think the reader should be coming at this with the concepts of TE link
and LSR in their heads so that the mapping is clear.
Ok, chalk this one (and probably the previous one) up to me not being a
GMPLS expert.
Ramon> no changes
Last paragraph on p.16: "trnaponders" -> "transponders". Also, I saw
"transceivers" earlier - if that's the same concept, only one term
should be used.
While "transponder" is technically correct, using "transceiver" would be more
consistent.
Ramon> Changed to transceiver, thanks. Probably the typo caused it being
missed in a S&R :)
Ok.
p.21, 1st para:
messages, and a specific frequency slot can be requeste on any
s/requeste/requested
Ramon> changed, thanks.
p.21:
In GMPLS the requested effective frequency slot is represented to the
TSpec present in the Path message, and the effective frequency slot
is mapped to the FlowSpec carried in the Resv message.
I believe those are RSVP-TE messages - that should be stated.
Ramon> Changed to "present in the RSVP-TE Path message (...) RSVP-TE
Resv message.
Thanks
p. 22:
d. n can change, but m needs to remain the same along the path.
This ensures that the effective frequency slot remains valid, but
allows the frequency slot to be moved within the spectrum from
hop to hop.
In full generality, that may require the ability to shift or convert a
frequency slot, which is a concept that doesn't appear to occur in the
draft prior to this point.
Penultimate paragraph of page 21.
Ok.
Ramon> no changes.
Figures 15 and 16 need their variables (e.g., m_a, FSb) somehow
labelled or explained
Ramon> Changed to
C B A
|Path(m_req) | ^ |
|---------> | # |
| | # ^
-^--------------^----------------#----------------#--
Effective # # # #
FS n, m # . . . . . . .#. . . . . . . . # . . . . . . . .# <-fixed
# # # # n
-v--------------v----------------#----------------#---
| | # v
| | # Resv |
| | v <------ |
| | |FlowSpec(n, m_a)|
| | <--------| |
| | FlowSpec (n, |
<--------| min(m_a, m_b))
FlowSpec (n, |
min(m_a, m_b, m_c))
m_a, m_b, m_c: Selected frequency slot widths
and
C B A
|Path(m_req) ^ | |
|---------> # | |
| # ^ ^
-^-------------#----------------#-----------------#--------
Effective # # # #
FS n, m # # # #
# # # #
-v-------------v----------------#-----------------#--------
| | # v
| | # Resv |
| | v <------ |
| | |FlowSpec(n_a, m_a)
| | <--------| |
| | FlowSpec (FSb [intersect] FSa)
<--------|
FlowSpec ([intersect] FSa,FSb,FSc)
n_a: Selected nominal central frequencyfr by node A
m_a: Selected frequency slot widths by node A
FSa, FSb, FSc: Frequency slot at each hop A, B, C
After Figure 16, the switch to the EFS acronym is a surprise, given
the extensive prior usage of the spelled-out term. This paragraph
contains all uses of the EFS acronym - I suggest removing that acronym
and spelling out the term.
Ramon> done.
Section 4.6: I don't understand why this sentence is in the middle of
the paragraph - it doesn't seem to describe an example of different
slot width granularities:
Consider a node with an application where the nominal
central frequency granularity is 12.5 GHz and where slot widths are
multiples of 25 GHz.
I'd suggest removing it.
Ramon> although sentence did try (badly) to describe a different slot
widths (by default the values are 6.25 and 12.5) the sentence did not
add value other than being a trivial example. Removed
5.1.1. What is L-band? This is the first time it's mentioned.
Ramon> changed to
The control plane architecture SHOULD allow for the support of
L-band (the wavelength range 1565 nm to 1625 nm) and S-band (the
wavelength range 1460 nm to 1530 nm).
also added
(...) the entire C-band (the wavelength range 1530–1565 nm, which
corresponds to the amplification range of erbium doped fiber
amplifiers)
Many thanks
Ramon