RE: Working Group last call on Four Ethernet Drafts

["PAPADIMITRIOU Dimitri" <[e-mail]>]

 Benoit:

> -----Original Message-----
> From: Benoit C Tremblay [mailto:[e-mail]] 
> Sent: Friday, January 30, 2009 5:54 PM
> To: PAPADIMITRIOU Dimitri; Nic Neate; [e-mail]
> Subject: RE: Working Group last call on Four Ethernet Drafts
> 
> Hello,
> 
> I would like to understand the motivation behind the use of the
> classtype object instead of simply using the index field as a 
> bit vector
> for ethernet priorities as it is specified in the OIF UNI 2.0.  This
> simplifies the interpretation of the index field and reduces 
> the number
> of objects that need to be included in the message.  

Because, more than 8 values are needed. Now if you only use 8 values the
current spec does not prevent to do so.

Thanks,
-dimitri.

> Regards,
> 
> Benoit Tremblay
> 
> > -----Original Message-----
> > From: [e-mail] 
> > [mailto:[e-mail]] On Behalf Of PAPADIMITRIOU Dimitri
> > Sent: January-30-09 8:42 AM
> > To: Nic Neate; [e-mail]
> > Subject: RE: Working Group last call on Four Ethernet Drafts
> > 
> > Nic:
> >  
> > > -----Original Message-----
> > > From: Nic Neate
> > > Sent: 29 January 2009 16:05
> > > To: [e-mail]
> > > Subject: RE: Working Group last call on Four Ethernet Drafts
> > > 
> > > Some low-level comments on
> > > draft-ietf-ccamp-ethernet-traffic-parameters-06.
> > > 
> > > 
> > > (1) TLV type definition (page 5):
> > > 
> > >               Type     Length   Format            Description 
> > >               
> > ------------------------------------------------------  
> > >               1        TBD      Reserved          Reserved value 
> > >               2        24       see Section 3.1   Ethernet 
> > Bandwidth 
> > >                                                   Profile 
> [MEF10.1] 
> > >               3        8        [GMPLS-ESVCS]     Layer 2 Control 
> > >                                                   
> Processing (L2CP)
> > > 
> > > The reference to section 3.1 should be 4.1.
> > 
> > indeed.
> > 
> > > (2) Ethernet Bandwidth Profile TLV Index field (page 6):
> > > 
> > >                       The index field value MUST correspond to at 
> > > least one
> > >            of the index value included in the extended ClassType 
> > > object ([DS-
> > >            TE], [MCOS]). 
> > > 
> > > The [DS-TE] reference is missing from chapter 11.  Note also that 
> > > CLASSTYPE is defined in [DS-TE] (RFC 4124) whereas 
> > EXTENDED_CLASSTYPE 
> > > is defined in [MCOS] (draft-minei-diffserv-te-multi-class).  This 
> > > could be made clearer by updating the text to say "The 
> index field 
> > > value MUST correspond to at least one of the index values 
> > included in 
> > > the CLASSTYPE object ([DS-TE]) or EXTENDED_CLASSTYPE object 
> > ([MCOS])".
> > 
> > I will clarify.
> > 
> > > (3) Ethernet Bandwidth Profile TLV Index field (page 6):
> > > 
> > >            A given index value j can be associated to at most N 
> > > Class-Type
> > >            values CTi (i =< N) of the extended Class-Type 
> > object. This
> > >            applies in case a set of one or more CTi maps a single 
> > > (shared) BW
> > >            profile. An example of value setting consists then in 
> > > assigning
> > >            an arbitrary value (comprised within the range 
> > [0x08,0xF8],
> > >            defined by the 5 MSB of the Index field) 
> associated to a 
> > > set of
> > >            CTi. This allows mapping to one of 248 pre-defined CTi 
> > > sets.
> > > 
> > > It is unclear exactly what the range of values for defining 
> > a set of 
> > > CTi is.
> > > -  The range [0x08,0xF8] contains 241 values, not 248.
> > > -  The 5 most significant bits of the Index field can only 
> > be used to 
> > > create 31 values (excluding 0), not 248.
> > 
> > There are 16 bits in this field, the 3 lsb are used to encode 
> > the 8 cti defined in DS-TE: from 0x00000|000 (0x00) to 
> > 0x00000|111 (0x07), so 8 values [0x00,0x08[ in the group 0.
> > 
> > The remaining values [0x08,0xFF] are allocated as follows:
> > 
> > - The values in the range [0x08,0xF8[ are used for encoding 
> > CTi sets so 240 sets (30 groups as defined per msb of 8 
> values each) 
> > 
> > - The values in the range [0xF8,0xFF] are reserved (= 8 sets) 
> > i.e. the last group of 8 values is reserved.
> > 
> > So, the last sentence should read "This allows mapping to one 
> > of 248 pre-defined CTi sets" but i will change the 
> > parenthesis from ] to [.
> > thanks for catching it. i also need to change capital to 
> > minus letters for msb and lsb description.
> > 
> > Thanks,
> > -dimitri.
> > 
> > > I think the text should say "...an arbirary value from the range 
> > > [0x08,0xFF] associated to a set of CTi.  This allows 
> > mapping to one of 
> > > 248 pre-defined CTI sets."
> > > 
> > > 
> > > Nic
> > > 
> > > 
> > > 
> > > -----Original Message-----
> > > From: [e-mail]
> > > [mailto:[e-mail]] On Behalf Of Aria - 
> Adrian Farrel 
> > > Personal
> > > Sent: 24 January 2009 22:27
> > > To: [e-mail]
> > > Cc: Brungard, Deborah A, ALABS
> > > Subject: Working Group last call on Four Ethernet Drafts
> > > 
> > > We want to last call four Ethernet drafts in a batch.
> > > 
> > > All of the drafts are quite short, but we need to give 
> you a little 
> > > extra time to read and comment, so this will be a three 
> > week last call 
> > > ending on Monday 16th February at 12 noon GMT.
> > > 
> > > Please send your comments to the CCAMP list.
> > > 
> > > The drafts are:
> > > 
> > > Ethernet Traffic Parameters
> > > draft-ietf-ccamp-ethernet-traffic-parameters-06.txt
> > > 
> > > Generalized MPLS (GMPLS) Data Channel Switching Capable
> > > (DCSC) and Channel Set Label Extensions 
> > > draft-ietf-ccamp-gmpls-dcsc-channel-ext-00.txt
> > > 
> > > Generalized MPLS (GMPLS) Support For Metro Ethernet Forum and
> > > G.8011 Ethernet Service Switching
> > > draft-ietf-ccamp-gmpls-ether-svcs-02.txt
> > > 
> > > Generalized MPLS (GMPLS) Support For Metro Ethernet Forum and
> > > G.8011 User-Network Interface (UNI)
> > > draft-ietf-ccamp-gmpls-mef-uni-01.txt
> > > 
> > > I note that the last three of these are due to expire on 
> > February 8th 
> > > (i.e.
> > > during the last call) so I hope that the authors can refresh the 
> > > drafts with a new revision with no other changes except 
> for the IPR 
> > > boilerplate.
> > > 
> > > Thanks,
> > > Adrian
> > > 
> > > 
> > > 
> > > 
> > 
> > 
>