Re: [MIB-DOCTORS] AD review of draft-ietf-adslmib-vdsl2-06.txt

["David Harrington" <[e-mail]>]

Wow! You need to take a closer look at this module.
In the past I just looked at your comments and made some suggestions.
(I was surprised to see a response to comments I forgot I made.)
I have not reviewed this module in depth. But their responses made me
look a little closer to understand what they were doing.

Here are a few of my favorites:

1) Xdsl2LineEntry  ::=
   SEQUENCE {
      xdsl2LineCnfgTemplate            SnmpAdminString,
      xdsl2LineCnfgFallbackTemplate    SnmpAdminString,
      xdsl2LineAlarmCnfgTemplate       SnmpAdminString,
      xdsl2LineCmndConfPmsf            Xdsl2ConfPmsForce,
      xdsl2LineCmndConfLdsf            Xdsl2LineLdsf,
      xdsl2LineCmndConfLdsfFailReason  Xdsl2LdsfResult,
      xdsl2LineCmndConfBpsc            Xdsl2LineBpsc,
      xdsl2LineCmndConfBpscFailReason  Xdsl2BpscResult,
      xdsl2LineCmndConfBpscReqCount    Unsigned32,
      xdsl2LineCmndAutomodeColdStart   TruthValue,
      xdsl2LineCmndConfReset           Xdsl2LineReset,
      xdsl2LineStatusActTemplate       SnmpAdminString,
      xdsl2LineStatusXtuTransSys       Xdsl2TransmissionModeType,
      xdsl2LineStatusPwrMngState       Xdsl2PowerMngState,
      xdsl2LineStatusInitResult        Xdsl2InitResult,
      xdsl2LineStatusLastStateDs       Xdsl2LastTransmittedState,
      xdsl2LineStatusLastStateUs       Xdsl2LastTransmittedState,
      xdsl2LineStatusXtur              Xdsl2LineStatus,
      xdsl2LineStatusXtuc              Xdsl2LineStatus,
      xdsl2LineStatusAttainableRateDs  Unsigned32,
      xdsl2LineStatusAttainableRateUs  Unsigned32,
      xdsl2LineStatusActPsdDs          Integer32,
      xdsl2LineStatusActPsdUs          Integer32,
      xdsl2LineStatusActAtpDs          Integer32,
      xdsl2LineStatusActAtpUs          Integer32,
      xdsl2LineStatusActProfile        Xdsl2LineProfiles,
      xdsl2LineStatusActLimitMask      Xdsl2LineLimitMask,
      xdsl2LineStatusActUs0Mask        Xdsl2LineUs0Mask,
      xdsl2LineStatusActSnrModeDs      Xdsl2LineSnrMode,
      xdsl2LineStatusActSnrModeUs      Xdsl2LineSnrMode,
      xdsl2LineStatusElectricalLength  Unsigned32,
      xdsl2LineStatusTssiDs            Xdsl2Tssi,
      xdsl2LineStatusTssiUs            Xdsl2Tssi,
      xdsl2LineStatusMrefPsdDs         Xdsl2MrefPsdDs,
      xdsl2LineStatusMrefPsdUs         Xdsl2MrefPsdUs,
      xdsl2LineStatusTrellisDs         TruthValue,
      xdsl2LineStatusTrellisUs         TruthValue,
      xdsl2LineStatusActualCe          Unsigned32
   }
I love the abrreviations everywhere.
Check out inconsistent abbreviations, such as
xdsl2LineCnfgFallbackTemplate which points to the
xdsl2LineConfTemplateTable. (cnfg and conf both mean configuration)
I note that RFC2662 uses "conf" consistently.

2) I especially like the "you're screwed" counter - the
xdsl2LineCmndConfBpscReqCount let's management station M1 know when
another management station M2 requested a measurement, which caused
the relevant counter to reset, such that M1's measurement 
is no longer valid. 
      "SNMP managers can use this attribute to check that the
       measurement results retrieved by the manager where not 
       interupted by another measurement request."

I thought we discouraged resetting counters in IETF MIB modules?
Has this MIB been seriously reviewed by a MIB technical advisor?
Has this MIB module been reviewed by any management application
designers?

3) note the naming conventions for counters are not being followed -
xdsl2LineCmndConfBpscReqCount should probably be
xdsl2LineCmndConfBpscRequests.

4) xdsl2LineCmndConfBpscReqCount - note that this counter is not a
Counter32, but an Unsigned32.

5) I am a bit concerned about the table that actually contains the
measurement - the xdsl2LineSegmentTable  The actual measurement is in
xdsl2LineSegmentBitsAlloc. It appears the value is overloaded to
provide subcarrier allocation status, as well as the result of a line
diagnostic test. 

6) from security consdierations "Unauthorized changes to
xdsl2LineCmndConfBpsc could cause an unscheduled bits per subcarrier
measurement to be carried out on the line.", and "Unauthorized changes
to xdsl2LineCmndConfLdsf could cause an unscheduled line test to be
carried out on the line." Since these two measurements report using
the same object, and since one measurement can interfere with the
validity of the measurement requested by a different manager, this
seems to have serious issues of information reliability.

7) On a larger scale, I am concerned that this appears to be a
**competing** standard to RFC2662. This does not complement, or
update, or obsolete RFC2662.

I have not yet gone into the two mib modules to determine how
compatible they are, but section 4 in this document makes me really
uneasy as a (previous) management app designer. Can a device support
both mib modules simultaneously? Can a manageent app use both MIB
modules simultaneously to manage the same ADSL functions on a device?

I note that rfc2662 does support using the entity mib, while this mib
module does not. Which makes me wonder about the compatibility between
the two mib modules.

8) Per section 4, you need to go read G.997.1 to understand the
conformance requirements for this MIB module. G.997.1 appears to not
be a free document.

--
I am NOT volunteering to be the MIB Doctor for this document. I have
dealt with huge MIB modules before, especially ones designed by people
from another SDO (e.g., Printer MIB). They are just way too painful to
work though, with people resistant to changes, and ADSL/VDSL is
totally unrelated to my day job.

But I think you need to have this MIB reviewed thoroughly, because it
looks to me like the designers do not have a good understanding of how
MIB modules should be designed for
interoperability with multiple simultaneous management apps, and for
usebaility by operators.

dbh


> -----Original Message-----
> From: [e-mail] 
> [mailto:[e-mail]] On Behalf Of Romascanu, 
> Dan (Dan)
> Sent: Thursday, January 22, 2009 9:38 AM
> To: Moti Morgenstern; David B Harrington
> Cc: MIB Doctors (E-mail); [e-mail]; Menachem Dodge; 
> [e-mail]; [e-mail]
> Subject: Re: [MIB-DOCTORS] AD review of 
> draft-ietf-adslmib-vdsl2-06.txt
> 
> Hi Moti,
> 
> Your proposed resolution to the comments looks satisfactory to me. 
> 
> Please proceed with creating a revised version of this document
which
> could be sent to IETF LC. 
> 
> Thanks and Regards,
> 
> Dan
>  
> 
> > -----Original Message-----
> > From: Moti Morgenstern [mailto:[e-mail]] 
> > Sent: Sunday, November 09, 2008 4:53 PM
> > To: Romascanu, Dan (Dan); David B Harrington
> > Cc: MIB Doctors (E-mail); [e-mail]; Menachem Dodge; 
> > [e-mail]; [e-mail]
> > Subject: RE: AD review of draft-ietf-adslmib-vdsl2-06.txt
> > 
> > Hi Dan and David,
> > 
> > Thank you for the comments regarding to 
> > draft-ietf-adslmib-vdsl2-06.txt.
> > We, the editors, reviewed the comments and here is our 
> > response to each (Dan's comments and then David's):
> > 
> > Comment T1
> > --------------
> > Comment suggests writing specifically in the MIB module, in 
> > the DESCRIPTION clauses of the MIB tables, what is the 
> > persistence policy relative to the objects in the table.
> > Response: We have a section in the document called "2.5.  
> > Persistence" which lists all persistent objects. If you 
> > search for each of these objects in the MIB, not all of the 
> > objects mention persistence in the DESCRIPTION clause. This 
> > is when the persistency is relevant to ALL objects and, when 
> > that happens, the DESCRIPTION clause of the containing table 
> > mentions: "Entries in this table MUST be maintained in a 
> > persistent manner." So we believe the current situation is 
> > satisfactory.
> > 
> > Comment T2
> > --------------
> > Comment claims that indexation by name is very resource consuming.
> > Response: The observation is basically correct. Despite that, 
> > we chose name indexes on purpose for ease of use by the 
> > operator and consistency with all other xDSL MIBs. 
> > 
> > Comment T3
> > --------------
> > Comment refers to rate-limiting notifications mentioned in 
> > the document.
> > Response: We will remove the rate limiting comment altogether 
> > because no other xDSL MIB mentions this.
> > 
> > Comment T4
> > --------------
> > Comment is a request to explain the Reserved bits in 
> > Xdsl2TransmissionModeType.
> > Response: We used reserved bits because we want the VDSL2 MIB 
> > to closely match the source document G.997.1.
> > 
> > Comment T5
> > --------------
> > Comment about the Textual Conventions with enumerated INTEGER 
> > syntax in which the list of values starts from zero.
> > Response: Similar to T4 above this was done on purpose. We 
> > want the MIB to closely match the source document G.997.1 
> > which already defines the range of enumerated values for many 
> > attributes. Since the syntax allows that we don't see it 
> > necessary to explain such considerations in the MIB. 
> > 
> > Comment T6
> > --------------
> > Comment about the Boolean objects with new TCs instead of 
> > using TruthValue
> > Response: We use TruthValue for 10 objects in the MIB, so, 
> > it's not that we are not familiar with that TC. The real 
> > reason, again, is that G.997.1 specifically defines the range 
> > of values and meaning for many attributes and our MIB follows 
> > it as much as we could. 
> > 
> > Taking the Textual Conventions mentioned in this comment as 
> examples:
> > Xdsl2LineLdsf- G.997.1 uses the value 0 to inhibit LDSF and 1 
> > for forcing LDSF. We preferred explicitly using those values. 
> > Xdsl2LineCeFlag- CEFLAG is defined by G.997.1 as a bitmap 
> > with a single bit. So, the values range is 0 and 1.
> > Xdsl2LineSnrMode-TruthValue uses the value 1 for true and 2 
> > for false, while G.997.1 uses 1 for disable and 2 for enable.
> > 
> > There is still one TC (Xdsl2LineForceInp) that could be 
> > replaced by the TruthValue. This will be performed.  
> > 
> > Comment T7
> > --------------
> > Comment regarding to using RowStatus only to delete rows
> > Response: We'll add text in the relevant DESCRIPTION clauses 
> > to explain that the agent rejects any SET operation which 
> > sets row status to a value other than delete(6)
> > 
> > Comment T8
> > --------------
> > Comment: why not replace inventory information in 
> > Xdsl2LineInventoryEntry with entPhysicalTable in RFC4133
> > Response: There are several arguments against the idea. 
> > However, the most important one is that we need to provide 
> > the inventory for both near-end and far-end units and that 
> > means two rows per each ifIndex. So, we cannot use RFC4133 
> > for inventory because the indexing scheme is inappropriate.
> > 
> > Comment T9
> > --------------
> > Comment: what happens when row is validated, etc.
> > Response: We agree that there's a need to perform several 
> > changes in the DESCRIPTION clause of row-status objects. This 
> > will include removing the text about validating the row 
> > content (i.e., as we cannot explain how we expect the 
> > implementation to perform that) as well as giving more 
> > information regarding to deleting a "referenced" row. 
> > 
> > Comment T10
> > --------------
> > Comment requests explanation about how the thresholds in the 
> > xdsl2ChAlarmConfProfileTable are being applied
> > Response: First of all, RFC3705 includes TCs and we use one 
> > of them in the referred table. Second, we suggest to add a 
> > text in the top part of the document explaining that only one 
> > threshold trap per interval will be issued once the 
> > individual counter exceeds the threshold defined for it. Is 
> > that acceptable?
> > 
> > Comment T11
> > --------------
> > Comment suggests that the compliance should be function of 
> > the flavor(s) of xDSL an agent implements. 
> > Response: The MODULE-COMPLIANCE section applies to all 
> > technology types: VDSL2/ADSL/ADSL2 and ADSL2+. The assumption 
> > of G.997.1 is that a typical implementation supports a 
> > mixture of technologies and transmission modes, not a single 
> > selected one, and that's why the abstract attribute was 
> > originally defined as a bit map rather than an enumeration of 
> > modes. The idea is that a typical operator wishes to serve 
> > all transmission modes supported by the remote units so he 
> > prepares, for each service package, a setup that can satisfy 
> > multiple use cases. 
> > Theoretically, Most of the MIB attributes are relevant for 
> > all or at least multiple xDSL technologies. There are still 
> > several which are relevant for a specific technology. We 
> > didn't want to complicate the compliance section because 
> > otherwise we would also need to handle attributes that are 
> > sometimes mandatory and sometimes optional (and what happens 
> > when you have a mixed scenario..?). So, we believe an agent 
> > implementation should conform to the compliance clause but 
> > also consult with ITU-T G.997.1 to see exactly what is 
> > required depending on the mixture of the various transmission 
> > modes relevant for the specific operator.  
> > 
> > Comment E1
> > --------------
> > Comment: s/ a more comprehensive/a complementary/
> > Response: The ADSL2 MIB is recommended for implementations 
> > that do not implement VDSL2 and do not want to be affected by 
> > it at all. The NG-xDSL MIB (this draft) has a wider scope of 
> > technologies and is recommended for implementations of VDSL2 
> > (only VDSL2 or together with older DSL technologies) and are 
> > ready to pay the "price". The "price" is expressed in various 
> > ways, e.g., the need to implement the "subcarrier group" 
> > related attributes for DELT also for ADSL/2/2+, although 
> > those attributes are meaningful only in VDSL2.
> > So the new MIB is not a complementary one but a solution for 
> > the same problem space, only with a wider set of DSL technologies.
> > 
> > Comment E2
> > --------------
> > Comment: list of abbreviations
> > Response: OK
> > 
> > Comment E3
> > --------------
> > Comment: Section 2.3.2 duplicates text from the TC-MIB module
itself
> > Response: OK. We'll delete it.
> > 
> > Comment E4
> > --------------
> > Comment: page 28 - s/This MIB allows supporting/This MIB 
> > module allows supporting
> > Response: OK.
> > 
> > Comment E5
> > --------------
> > Comment: page 31 - s/ Network Elements may /Network Elements 
> > MAY/ and s/The xTU-C should / The xTU-C SHOULD/
> > Response: OK.
> > 
> > Comment E6
> > --------------
> > Comment: In SMIv2 we do not usually speak about attributes 
> > but about objects
> > Response: We'll correct the relevant text.
> > 
> > Comment E7
> > --------------
> > Comment: DESCRIPTION clause of the MIB module duplicates text 
> > in the Overview section
> > Response: We'll use a short summary in the overview section 
> > and a full description in the MIB module.
> > 
> > Comment E8
> > --------------
> > Comment: 'an ifType of vdsl(xxx)'. It is good to add at least 
> > at the first occurrence a note to the RFC Editor
> > Response: OK.
> > 
> > Comment E9
> > --------------
> > Comment: The DESCRIPTION clauses of Valid Intervals and 
> > Invalid Intervals are almost void
> > Response: We'll fill the DESCRIPTION clauses for these objects
> > 
> > Comment E10
> > --------------
> > Comment: Section 4 would benefit from one or two examples of 
> > configuration templates
> > Response: We prefer not to add any additional tutorial 
> > material into the VDSL2 draft MIB. 
> > 
> > 
> > Comments from David B Harrington
> > --------------------------------
> > 
> > Comment: "One suggestion would be to put the 
> > VDSL2-LINE-TC-MIB in a separate document."
> > Response: This would require 2 RFC numbers for VDSL2 and we 
> > don't like it. We also have an experience from RFC4706 and 
> > the draft ANCP AN MIB (draft-ietf-ancp-mib-an-03) that have 
> > the same structure as this draft.
> > 
> > Comment:"I strongly support Dan's request to eliminate some 
> > of the redundancy in this document, and I recommend starting 
> > with the textual convention triplication. You can probably 
> > save 30% of the document length getting rid of unnecessary 
> > textual convention definitions."
> > Response: We have no real objection to removing the TCs that 
> > are referenced only once and placing the definition in the 
> > object definition. Is that also agreed by Dan? Any comment 
> > from the WG members?
> > 
> > Comment: "It might be good to have a manageability 
> > considerations section that discusses using short strings and 
> > internal choices to minimize the resource impact."
> > Response: We don't think this is necessary.
> > 
> > Dan and David, Thanks again for your time and your comments.
> > 
> > On behalf of the co-editors,
> > Moti Morgenstern
> > 
> _______________________________________________
> MIB-DOCTORS mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/mib-doctors
> 

_______________________________________________
MIB-DOCTORS mailing list
[e-mail]
https://www.ietf.org/mailman/listinfo/mib-doctors