Definition of Managed Objects for the Neighborhood Discovery Protocol
draft-ietf-manet-nhdp-mib-19

[Ballot comment]
Thank you for your good work, which addresses my DISCUSS/COMMENT

Note regarding the applicability statement: This is solved, as we discussed, but I'll keep this little sentence in one
corner of my head "A fuller discussion of MANET network management use cases
and challenges will be provided elsewhere."

[Ballot comment]
- Abstract

  This document defines a portion of the Management Information Base
  (MIB) for use with network management protocols in the Internet
  community.  In particular, it describes objects for configuring
  parameters of the Neighborhood Discovery Protocol (NHDP) process on a
  router. 

Should this be "MANET router"?
"MANET router" is used in RFC6130
Alternatively, I see in this document:
"NHDP router"
"router in a Mobile Ad Hoc Network (MANET)"
Same remark for the introduction, which uses the same text.

[Ballot discuss]
[Updated the DISCUSS text, to be more precise, to summarize the discussion so far, and include actionable points]

1. The OPS-DIR Review by Tom Nadeau raises some significant issues. 
As far as I can tell, all issues are/will be addressed. Waiting for the next version of the draft to clear this part.

2. While reading through the document, I've been really waiting for an applicability statement section on how MANET routers based network will be managed? Note that the write-up mentions: "This document shepherd is not aware of existing implementations of this MIB although several implementers have indicated interest in the work."
So the potential implementers could share their views

I'm really after 3 different parts in the applicability statement (or call it "use cases" if you prefer): configuration management, monitoring, and notification. And I have questions such as:
* one static NMS or each MANET router configures/monitors his (newly attached) neighbor(s)?
* where are the notifications sent to? To a neighbor or the static NMS?
* Do you expect all MANET routers to always have connectivity to a static NMS?
* etc...
Obviously, monitoring, for example, is important in ad-hoc networks, but please let us know how you envision doing this. For example, right now, I see the term "appropriate SNMP management stations", and I have no clue what's "appropriate" in the management of an ad-hoc network.

I see some other MANET MIB modules...
* draft-ietf-manet-olsrv2-mib-04    Definition of Managed Objects for the Optimized Link State Routing Protocol       
* draft-ietf-manet-report-mib-02    Definition of Managed Objects for Performance Reporting
* draft-ietf-manet-smf-mib-04        Definition of Managed Objects for the Manet Simplified Multicast Framework Relay 
                                                Set
However, this document is the first MANET MIB module to reach the IESG, so those operational questions are important.
Maybe this part of the DISCUSS will be solved by a new WG document: MANET manageability considerations

[Ballot discuss]
- There is a clear DISCUSS coming from the MIB-doctor review. Please address it.

- While reading through the document, I've been really waiting for an applicability statement section.
A sentence [RFC6130] such as "if router A is able to exchange packets with router B, and router B is able to exchange packets with router C, this does not guarantee that router A and router C can exchange packets directly" got me thinking about the management and operational aspects. How do you plan on using this MIB module, taking into account that you're dealing with an ad-hoc network?

Note that the write-up mentions: "This document shepherd is not aware of existing implementations of this MIB although several implementers have indicated interest in the work."
So the potential implementers should have some views on the following questions.

I'm really after 3 different parts in the applicability statement (or call it use cases if you want to): configuration management, monitoring, and notification. And I have questions such as:
* one static NMS or each MANET router configures/monitors his (newly attached) neighbor(s)?
* where are the notifications sent to?
* Do you expect all MANET routers to always have connectivity to a static NMS?
* etc...
Obviously, monitoring, for example, is important in ad-hoc networks, but please let us know how you envision doing this. For example, right now, I see the term "appropriate SNMP management stations", and I have no clue what's "appropriate" in the management of an ad-hoc network.

I see some other MANET MIB modules...
* draft-ietf-manet-olsrv2-mib-04 Definition of Managed Objects for the Optimized Link State Routing Protocol
* draft-ietf-manet-report-mib-02 Definition of Managed Objects for Performance Reporting
* draft-ietf-manet-smf-mib-04 Definition of Managed Objects for the Manet Simplified Multicast Framework Relay Set
However, this document is the first MANET MIB module to reach the IESG, so those operational questions are important.
If this is expressed in a different document, let me know, and I can review it.

Along the same lines,
1. see Al Morton's review, part of the OPS-DIR
> General question: Is a lossy, mobile ad-hoc network compatible with
> SNMPv3? In other words, will the link performance information needed
> for troubleshooting really be available when it is needed most, and
> critical nodes may be unreachable?
2. Read Ron's Bonica's Abstain

[Ballot comment]
- Abstract

  This document defines a portion of the Management Information Base
  (MIB) for use with network management protocols in the Internet
  community.  In particular, it describes objects for configuring
  parameters of the Neighborhood Discovery Protocol (NHDP) process on a
  router. 

Should this be "MANET router"?
"MANET router" is used in RFC6130
Alternatively, I see in this document:
"NHDP router"
"router in a Mobile Ad Hoc Network (MANET)"
Same remark for the introduction, which uses the same text.

- First occurence of NHDP must be expanded.

[Ballot comment]
This document fails to provide use cases. Because it doesn't provide use cases, it is not clear that any/all of its objects are useful.

While this comment might be leveled concerning regarding MIB, it is particularly applicable in this case because we have so little operational experience with ad hoc networks. Will the ad hoc network include a NOC? What policy will that NOC attempt to enforce? What information/control will the NOC need to enforce that policy.

[Ballot comment]
I am voting no objection on the basis that other members of the IESG who are more knowledgeable on MIBS will have reviewed this.

[Ballot comment]
Dear all,

I've been unable to find a volunteered MIB-doctor for this draft. Therefore, I have to defer this draft.
Actually, Adrian did the right thing by being proactive with the MIB doctors, at the time of the IETF last-call
My bad for not following up, and assigning someone. I will pay more attention from now on...

Regards, Benoit.

[Ballot discuss]
1) s5: I think the 3rd to last and 2nd to last paragraphs need to be changed (Benoit might be able to help with this) based on the following:

https://www.ietf.org/mail-archive/web/mib-doctors/current/msg01369.html

text is here (references are in the above link):

https://www.ietf.org/mail-archive/web/mib-doctors/current/msg01367.html

2) s5 contains the following:

  Therefore, when implementing these
  capabilities, the full use of SNMPv3 cryptographic mechanisms for
  authentication and privacy is RECOMMENDED.

There's a couple of ways to do this.  Is there a preferred mechanism?  The change in #1 gets to some of this, but does this use have a preference: TSM over SSH?  I ask because RFC3410 algs are really, really out of date (e.g., uses DES and I can't say it'd provide any privacy) and it'd be good to point to something better in light of the uses of this MIB.

IESG:

IANA has reviewed draft-ietf-manet-nhdp-mib-12.txt and has the following
comments:

IANA understands that, upon approval of this document, there is
a single action which IANA must complete.

In the SMI Network Management MGMT Codes Internet-standard MIBsubregistry
of the Network Management Parameters registry located at:

http://www.iana.org/assignments/smi-numbers

a new MIB will be registered as follows:

Decimal: [ TBD by IANA at time of registration ]
Name: nhdpMIB
Description: Neighborhood Discovery Protocol
References: [ RFC-to-be ]

IANA understands this to be the only action required of IANA upon
approval of this document.

Note: The actions requested in this document will not be completed until
the document has been approved for publication as an RFC. This message is
only to confirm what actions will be performed.

State changed to In Last Call from Last Call Requested

The following Last Call Announcement was sent out:

From: The IESG

To: IETF-Announce

CC:

Reply-To: ietf@ietf.org

Subject: Last Call:  (Definition of Managed Objects for the Neighborhood Discovery Protocol) to Proposed Standard





The IESG has received a request from the Mobile Ad-hoc Networks WG

(manet) to consider the following document:

- 'Definition of Managed Objects for the Neighborhood Discovery Protocol'

  as a Proposed Standard



The IESG plans to make a decision in the next few weeks, and solicits

final comments on this action. Please send substantive comments to the

ietf@ietf.org mailing lists by 2012-04-16. The span of this last call has

been extended to account for the IETF meeting. Exceptionally, comments

may be sent to iesg@ietf.org instead. In either case, please retain the

beginning of the Subject line to allow automated sorting.



Abstract



  This memo defines a portion of the Management Information Base (MIB)

  for use with network management protocols in the Internet community.

  In particular, it describes objects for configuring parameters of the

  Neighborhood Discovery Protocol (NHDP) process on a router.  The MIB

  module defined in this memo, denoted NHDP-MIB, also reports state,

  performance information and notifications.  This additional state and

  performance information is useful to troubleshoot problems and

  performance issues during neighbor discovery.





The file can be obtained via

http://datatracker.ietf.org/doc/draft-ietf-manet-nhdp-mib/



IESG discussion can be tracked via

http://datatracker.ietf.org/doc/draft-ietf-manet-nhdp-mib/ballot/





No IPR declarations have been submitted directly on this I-D.

AD review...

Hi,

Many apologies for a delayed AD review of this document. I love MIB
modules deeply, but I did want to clear out the other MANET documents
from the pipe before working on this one.

As usual, the purpose of my review is to catch and clean up issues that
might show later in the process, thereby saving community and IESG
review effort, and giving the document a smoother ride through the
system and a greater likelihood of approval.

This is a substantial document and is generally solid. The number of
comments reflects the fact that writing MIB modules is a real back-
breaker. I think it is important to note that I did not have any
concerns about the structure of the module or the tables (except one
small suggestion about moving the TCs into a separate module). Most
of the comments are at the level of MIB nits.

All comments are, of course, up for discussion and debate.

I think we need a new revision at this point. Then we can go to MIB
Doctor review and IETF last call at the same time.

Many thanks,
Adrian

---

5.1.3.1

  The majority of critical events occur when NHDP is enabled on a
  router, at which time the symmetric neighbors and two-hop neighbors
  of the NHDP router are discovered.

I think we can clarify this because there are currently two
interpretations:

- when NHDP is operational (enabled)
- when NHDP is initiated (enabled or the first time)

From context, I think you mean...

  The majority of critical events occur when NHDP is first enabled on a
  router, at which time the symmetric neighbors and two-hop neighbors
  of the NHDP router are discovered.

---

5.1.3.1

  To avoid unnecessary notifications, a router should not
  originate expected notifications until a certain time interval has
  elapsed, which is to be predefined by the network manager.

1. Is this "SHOULD NOT"?
2. Is there a recommended default for such a timer value? I think you
  should give one if you can.

---

5.1.3.2

  Appropriate values for the window time and upper bound are to be
  selected by the network manager and depend on the deployment of the
  MANET.

I believe you here, but I would welcome some more guidance for the
implementer and the deployer. Is there anything you can add?

---

5.1.3.3

OLD
  Similar to the according mechanism in [RFC4750], only one
  notification is sent per event.
NEW
  Similar to the mechanism in [RFC4750], only one notification is sent
  per event.
END

---

Section 6

  This section specifies the relationship of the MIB modules contained

s/modules/module/

---

You have included Section 6.3 which is perfect.
In the light of this, there is a body of opinion amongst "MIB experts"
that citation notation should not be used in the body of the MIB
module. This is because the module will be ripped from the document
for compilation and passed around as a stand-alone piece of text.

Thus, just take the square brackets out form the comments in the
IMPORT clauses.

Make the same change to any Description and Reference clauses.

---

Don't forget to make any necessary changes to the CONTACT-INFO clause.

---

The copyright date in the MODULE-IDENTITY clause is a little out-of-
date.

---

For sound reasons, revisions of the MIB module posted in Internet-Drafts
are not considered as "Revisions". That is, per the boilerplate, it is
inappropriate to use Internet-Drafts as reference material or to cite
them other than as "work in progress.

That means that the first revision of the MIB Module is the one that
gets published in the RFC.

Can you tidy the Revision clauses so that there is only one, and it
applies to the RFC-that-will-be. The RFC Editor will fix up the date on
publication.

---

NeighborIfIndex

I'm not saying what you have is wrong, but why did you decide to not
assign this TC the Syntax InterfaceIndex?

---

For nhdpInterfaceTable I think it is necessary to say that when the
corresponding entry with ifIndex value is deleted from the Interface
Table, the entry in this table is automatically deleted.

---

nhdpIfIndex has Syntax InterfaceIndexOrZero, but the Description says
"The ifIndex for this interface."

Note that according to RFC 2863, ifIndex has Syntax InterfaceIndex.

So you need to decide to either:
- change the Syntax of nhdpIfIndex
or
- add an explanation of what it means to have a zero value

---

nhdpIfStatus

I *think* it is more normal to write:

      DEFVAL { false(2) }

---

nhdpInitialPending

Given the constraint conditions described for this object, I don't
think it is appropriate to have a Default clause. That is, if the
constraining objects are not defaulted for whatever reason, then this
object cannot be allowed to default.

This is correctly reflected in you making the object read-only, and we
should note that there is no value in assigning a Default for a read-
only object.

---

Finding nhdpIfRowStatus and seeing the text talk about row creation, I
wondered why none of the objects in the row under the control of this
object have Max-Access read-create. What you have is legal, but it means
that the row is created with all the defaults in place and then can be
modified if the management application wants to.

Since there is no equivalent of an adminStatus, this means that for a
while the row (and associated protocol elements) will operate with the
default values rather than the values the management application wanted
to set.

---

The second paragraph of the description of nhdpLibLocalIfSetTable begins
"It consists of..."  This is mildly ambiguous because the previous
paragraph ends with a sentence about nhdpIfIndex.

---

nhdpLibLocalIfSetIpAddrType should contain a comment about the valid
values since I assume you don't support the full set defined for
InetAddressType. You can also constrain the interpretation of
nhdpLibLocalIfSetIpAddr according to nhdpLibLocalIfSetIpAddrType.

You would write...

  nhdpLibLocalIfSetIpAddrType  OBJECT-TYPE
      SYNTAX      InetAddressType
      MAX-ACCESS  read-write
      STATUS      current
      DESCRIPTION
        "The type of the nhdpLibLocalIfSetIpAddr
          in the InetAddress MIB [RFC4001].

          Only the values unknown(0), ipv4(1), and
          ipv6(2) are supported."
      REFERENCE
        "[RFC6130]."
  ::= { nhdpLibLocalIfSetEntry 1 }

  nhdpLibLocalIfSetIpAddr  OBJECT-TYPE
      SYNTAX      InetAddress
      MAX-ACCESS  read-write
      STATUS      current
      DESCRIPTION
        "nhdpLibLocalIfSetAddr is an
          address of an interface of
          this router.

          This object is interpretted according to
          the setting of nhdpLibLocalIfSetIpAddrType."
      REFERENCE
        "[RFC6130]."

This will also need to show up in the conformance statements. You
write something like...

OBJECT nhdpLibLocalIfSetIpAddrType
  SYNTAX      InetAddressType { unknown(0), ipv4(1), ipv6(2) }
  MIN-ACCESS  read-only
  DESCRIPTION  "Only unknown(0), ipv4(1), and ipv6(2) support
                is required."

And similarly...

OBJECT nhdpLibLocalIfSetIpAddr
  SYNTAX      InetAddress (SIZE(0|4|16))
  MIN-ACCESS  read-only
  DESCRIPTION "An implementation is only required to support
              address sizes of:
                0 for address type unknown(0)
                4 for address type ipv4(1)
                16 for address type ipv6(2)."

Lastly, is the value zero allowed for nhdpLibLocalIfSetIpAddr? If so
you will need some text in its DESCRIPTION clause like...

    If this object is set to 0, the value of
    nhdpLibLocalIfSetIpAddrType must be set to
    unknown(0)."

You would need to do this for all uses of InetAddress[Type]

---

You need to do a general search and replace s/MIB/MIB module/
There is only one MIB and this MIB module forms part of it.
For example...

nhdpStateObjGrp    OBJECT IDENTIFIER ::= { nhdpObjects 2 }

  -- Two new constructs have been defined in this MIB for
  -- indexing into the following
  -- tables and indexing into other tables in other MIBs.

---

Question...

nhdpStateObjGrp    OBJECT IDENTIFIER ::= { nhdpObjects 2 }

  -- Two new constructs have been defined in this MIB for
  -- indexing into the following
  -- tables and indexing into other tables in other MIBs.

1. This text seems to come very late in the module. How about
  relocating to be next to the definition of the "constructs"?

2. I think s/constructs/Textual Conventions/

3. If it is clear that you will use these TCs to index other
  MIB modules, you should seriously consider putting them in
  their own module. You can keep this new module in this
  document, but making it a separate module makes imports from
  other modules much more simple.

---

nhdpUpTime could use a display hint clause.

---

  nhdpInterfaceStateTable  OBJECT-TYPE
      SYNTAX      SEQUENCE OF NhdpInterfaceStateEntry
      MAX-ACCESS  not-accessible
      STATUS      current
      DESCRIPTION
        "nhdpInterfaceStateTable lists state information
          related to specific interfaces of this NHDP router.
          The ifIndex is from the interfaces group
          defined in the Interfaces Group MIB.

Unless I am mistaken, you are not using ifIndex, but ndhpIfIndex.
So you should say...

          The value of ndhpIfIndex is an ifIndex from the
          interfaces group defined in the Interfaces Group
          MIB.

Similarly nhdpInterfaceStateEntry (which is missing a REFERENCE clause.

---

There is nothing wrong, IMHO, with using TimeTicks for nhdpUpTime and
nhdpIfStateUpTime. However, this places a burden on the implementation
rather than the agent.

A common approach is to grab the sysUpTime value when the instance was
initialised, and to only report that. An agent that is interested in
the amount of time that the instance has been up, can read sysUpTime
and do the calculation itself.

If you grab sysUpTime, you use the Textual Convention TimeStamp which
ironically has the Syntax TimeTicks. The difference is that it is frozen
rather than your usage which is rolling.

---

As I understand it, nhdpIibLinkSetLTime is a time in the future. I'm not
sure, butI don't think you are allowed to use TimeStamp like this
because it is supposed to report on an event that has already happened.
This may be a bit pettifogging and we should maybe make a list of
questions to ask the MIB Doctor.

---

nhdpInterfacePerfTable etc.

For you to think about depending on the interface speeds you think you
are supporting. The requirements on the use of 32-bit and 64-bit
counters copied from RFC 2863 are as follows:

  For interfaces that operate at 20,000,000 (20 million) bits per
  second or less, 32-bit byte and packet counters MUST be supported.
  For interfaces that operate faster than 20,000,000 bits/second,
  and slower than 650,000,000 bits/second, 32-bit packet counters
  MUST be supported and 64-bit octet counters MUST be supported.
  For interfaces that operate at 650,000,000 bits/second or faster,
  64-bit packet counters AND 64-bit octet counters MUST be
  supported.

---

  nhdpSetNotification OBJECT-TYPE
          SYNTAX      OCTET STRING (SIZE(4))
          MAX-ACCESS  read-write
          STATUS      current
          DESCRIPTION
            "A 4-octet string serving as a bit map for
            the notification events defined by the NHDP
            notifications. This object is used to enable
            and disable specific NHDP notifications where
            a 1 in the bit field represents enabled. The
            right-most bit (least significant) represents
            notification 0.

            This object is persistent and when written
            the entity SHOULD save the change to
            non-volatile storage.
            "
          ::= { nhdpNotificationsControl 1 }

I find the use of Octet String in this way a bit icky. It is, of course,
possible to represent the whole MIB module using Octet Strings, but we
don't. Did you consider and reject the BITS construct? It seems neater
and more explicit.

I also spent some time trying to work out which notification is
"notification 0" because the first notification listed is
{ nhdpNotificationsObjects 1 }

---

Should the ChangeThreshold and ChangeWindow family of objects have
DEFVALs?

---

nhdpNbrState, nhdp2HopNbrState, and nhdpIfState are (correctly)
read-only. They shouldn't have DEFVALs defined because they report what
they report.

You might want to add to the Description...

If the state of foo is unknown, an implementation should return bar(n)

Or you might want to define specific "unknown" values to handle this,

---

Really good job on the Security Considerations section!

---

I would fold Sections 10 and 11 together since "Contributors" in an
RFC are effectively at "Author" status.

  (1.a) Stan Ratliff (sratliff@cisco.com) is the document shepherd
        for this document. The shepherd has personally reviewed the
        document, and believes it is ready for forwarding to the
        IESG for publication.

  (1.b) The document has had adequate review from both key working
        group members and from key non-WG members. The shepherd does
        not have any concerns about the depth or breadth of the
        reviews that have been performed.

  (1.c) The shepherd does not have any concerns about the document
        needing additional review.

  (1.d) The shepherd does not have any concerns or issues with the
        document that the responsible Area Director or the IESG
        need to be aware of. IPR disclosures were not necessary,
        therefore, none have been filed.

  (1.e) Working group consensus behind this document is solid. The
        document represents strong concurrence of the working group
        as a whole, the the WG understands and agrees with the
        document. 

  (1.f) No one has threatened appeal or has indicated discontent with
        the document.

  (1.g) The document shepherd has run the "idnits" tool against the
        document; some minor issues were detected and resolved. The
        document has met all required formal review criteria. The
        document will be reviewed by MIB Doctors as a function of
        submission to the IESG.

  (1.h) The document has split its references into normative and
        informative. To the shepherd's knowledge, there are no
        normative references to documents that are not ready for
        advancement or are in an unclear state. There are no
        downward references.

  (1.i) The shepherd has verified that document IANA consideration
        section exists, and is consistent with the body of the
        document. No protocol extensions are requested. The necessary
        IANA registries are clearly defined. No new registries are
        requested. As a function of moving the document to the IESG,
        we request a MIB Doctor review from the OPS AD.

  (1.j) The document has been run through the "smilint" checker.
        Warnings exist due to references to "FLOAT-TC-MIB", however,
        those references appear to be due to a deficiency of the
        tool (e.g., RFC6340 is 'too new' to be in view by the tool).

  (1.k) The IESG approval announcement includes a Document
        Announcement Write-Up. Please provide such a Document
        Announcement Write-Up? Recent examples can be found in the
        "Action" announcements for approved documents. The approval
        announcement contains the following sections:

    Technical Summary
        This document defines a portion of the Management Information
        Base (MIB) for use with the Neighborhood Discovery Protocol
        (NHDP) developed in the MANET working group.

    Working Group Summary
        The process for reaching working group consensus on this
        was smooth; no controversy existed. Working group consensus
        behind the document is solid.

    Document Quality
        This document shepherd is not aware of existing implementations
        of this MIB.