[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: GMPLS Last Calls - generalized-04/rsvp-te-03
Here are my comments/questions. I apologize if
I overlook the text that might already explain
my questions. A pointer will do in that case.
1. In section 3, it says that "generalized label does not
contain a type filed, instead the nodes are expected
to know from context what type of label to expect".
This principle is applied to all other label object
except LABEL_SET. Is there a particular reason to
include "label type" (section 3.5.1) in LABEL_SET object ?
2. Section 3.1.1, why does LSP Enc. Type includes "Packet",
"Ethernet V2/DIX", and "Ethernet 802.3" ? Aren't they all
packet and their label type can be derived from its link ?
If not, don't we also need ATM, FrameRelay in the code point,
in particular section 18.104.22.168 "Other Labels" talks about
ATM, FrameRelay labels.
I have doubt as to whether we really need LSP encoding type
since that is a service discovery issue and should
be known before signaling, in particular in the context
of RSVP where upstream selects the link. Can some one
explain the benefit of having it, besides from verifying
the consistency of the request and the link type ?
3. Section 3.1.2 bandwidth encoding. The text said that
these bandwidth encoding does not apply to SONET/SDH
technology. I believe they also do not apply to ANSI/ETSI
PDH technology, for the same reason we exclude SONET/SDH.
Isn't it more precise to say which technology the bandwidth
encoding applies to ?
4. Section 22.214.171.124 Port and Wavelength Labels. The description
of "Label" using the term "sender" and "receiver". In RSVP
these two terms have special meaning (upstream and downstream),
so I am not clear whether "sender" means the sender of the
(RESV) message or the "sender" of the LSP. This is particular
troublesome when it is referenced, such as "The information
carried in a suggested label is identical to a generalized
label". Suggest to use "upstream" or "downstream".
5. Section 3.3 Waveband Switching. How do we
distinguish 3 wavelength labels from waveband labels ?
Or is this always used in conjunction with "lambda LSP" ?
The text also said waveband Identifier is "The value is
selected by the sender and reused in all subsequent related
messages." What are the subsequent related message that
would include waveband identifier but not the waveband label
The same issue with using "sender" here. Suggest use
upstream or downstream.
6. Section 3.5.1 Required Information for Label Set.
Can each SubChannel x be a list of labels,
e.g. for an STS-12 each of them can include 12 SKLM
values ? So, if I have 24 time slot that can be used
for an STS-12 connection, I have to specify all combination
that are allowed, or can I just specify the 24 SKLM
values as subchannel 1 when requesting a STS-12 connection ?
If so, why do I need more than one "subchannels" ?
1. section 2.1.2. Does it mean one must accept both "control load"
and "guarantee service" in flowspec ?
2. section 5.1. Procedure for Explicit Label Control.
The text said that "the preceding subobject must be a strict object",
isn't it enough to say that "the preceding object must uniquely
identify a link ?"
> -----Original Message-----
> From: George Swallow [mailto:firstname.lastname@example.org]
> Sent: Monday, May 14, 2001 10:52 AM
> To: email@example.com; mpls@UU.NET
> Subject: GMPLS Last Calls
> This message initiates a last call on four GMPLS drafts. The last
> call is being issued jointly in the MPLS and CCAMP workgroups.
> The drafts are:
> 1. Generalized MPLS - Signaling Functional Description
> 2. Generalized MPLS Signaling - RSVP-TE Extensions
> 3. Generalized MPLS Signaling - CR-LDP Extensions
> 4. GMPLS Extensions for SONET and SDH Control
> The last call closes May 29, 12 pm GMT.
> - V2KG (Vijay, Vijay, Kireeti, & George)
> George Swallow Cisco Systems (978) 244-8143
> 250 Apollo Drive
> Chelmsford, Ma 01824