
From waldbusser@nextbeacon.com  Wed Aug 14 16:16:02 2013
Return-Path: <waldbusser@nextbeacon.com>
X-Original-To: mib-doctors@ietfa.amsl.com
Delivered-To: mib-doctors@ietfa.amsl.com
Received: from localhost (localhost [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 9036821F9C6A for <mib-doctors@ietfa.amsl.com>; Wed, 14 Aug 2013 16:16:02 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.598
X-Spam-Level: 
X-Spam-Status: No, score=-2.598 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, HTML_MESSAGE=0.001]
Received: from mail.ietf.org ([12.22.58.30]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id A7iUvQN2DT3l for <mib-doctors@ietfa.amsl.com>; Wed, 14 Aug 2013 16:15:56 -0700 (PDT)
Received: from mxout-08.mxes.net (mxout-08.mxes.net [216.86.168.183]) by ietfa.amsl.com (Postfix) with ESMTP id 6A2DC21F9223 for <mib-doctors@ietf.org>; Wed, 14 Aug 2013 16:15:56 -0700 (PDT)
Received: from [127.0.0.1] (unknown [24.6.159.106]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by smtp.mxes.net (Postfix) with ESMTPSA id 1E8E9509B5; Wed, 14 Aug 2013 19:15:51 -0400 (EDT)
Message-ID: <520C0FA7.1010904@nextbeacon.com>
Date: Wed, 14 Aug 2013 16:15:51 -0700
From: Steve Waldbusser <waldbusser@nextbeacon.com>
User-Agent: Mozilla/5.0 (Windows NT 5.1; rv:17.0) Gecko/20130801 Thunderbird/17.0.8
MIME-Version: 1.0
To: Agrahara Kiran Koushik <kkoushik@cisco.com>
References: <51E5ED0F.10207@nextbeacon.com> <A689AE1D-9EBC-4160-A5C3-E4A653900E21@cisco.com>
In-Reply-To: <A689AE1D-9EBC-4160-A5C3-E4A653900E21@cisco.com>
Content-Type: multipart/alternative; boundary="------------080809080104040308070905"
X-Antivirus: avast! (VPS 130814-1, 08/14/2013), Outbound message
X-Antivirus-Status: Clean
Cc: mib-doctors@ietf.org, "Giles Heron \(giheron\)" <giheron@cisco.com>, draft-ietf-l2vpn-vpls-mib.all@tools.ietf.org
Subject: Re: [MIB-DOCTORS] Review of draft-ietf-l2vpn-vpls-mib-10
X-BeenThere: mib-doctors@ietf.org
X-Mailman-Version: 2.1.12
Precedence: list
List-Id: MIB Doctors list <mib-doctors.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/mib-doctors>, <mailto:mib-doctors-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/mib-doctors>
List-Post: <mailto:mib-doctors@ietf.org>
List-Help: <mailto:mib-doctors-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/mib-doctors>, <mailto:mib-doctors-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 14 Aug 2013 23:16:02 -0000

This is a multi-part message in MIME format.
--------------080809080104040308070905
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

On 8/14/2013 2:07 PM, Agrahara Kiran Koushik wrote:
>
>> 8. This document is structured as three separate MIB modules. Did you consider putting everything in
>>    one MIB module and using the conformance section to modularize the MIB into 3 optional groups?
> [kiran] We had specific requests in the beginning to separate the MIB into different modules instead
> of one. We'd like to keep it that way if possible.
It's not my preferred choice but it's a valid choice - I don't have a 
big problem with it.

>
>> 12. Generally: How does the agent know that you intend to start BGP or LDP on a VPLS entry that you
>> have created with SNMP?
> [kiran] Not sure why it would be required?
Here's what I was thinking:
Let's say a row is created via SNMP in the vplsConfigTable. Everything 
the agent knows about that VPLS service is contained in that row. I'm 
assuming the agent will need to be told whether to use BGP or LDP for 
that service.

Perhaps it's a side-effect of whether you created an entry in the 
VplsLdpConfigTable or created one in the vplsBgpConfigTable? Could work 
I guess but it would be good to prohibit situations where neither entry 
is created or both are created (especially the latter). Most important 
would be to define authoritatively how the agent knows. It seemed like 
the other BDP and LDP tables could be empty even if the corresponding 
service was enabled so they wouldn't be good choices.

Thanks,
Steve


>
>
>
>
>
> Thanks
> Kiran
>
>
> On Jul 16, 2013, at 8:02 PM, Steve Waldbusser <waldbusser@nextbeacon.com> wrote:
>
>> Hello,
>>
>> Here's my review of the the VPLS MIB.
>>
>> 1. There are three tables that have a many-to-one relationship with the vplsConfigTable AND that
>>    allow SNMP create operations. A potential ambiguity exists in the creation process for these
>>    tables. For example one could create entries in the vplsPwBindTable that reference VPLS entries
>>    that don't yet exist: IOW one could create vplsPWBindRowStatus.VPLS-A.PW-B and THEN create vpls-A
>>    in the vplsConfigTable. It is perfectly alright for the inactive rows in the vplsPWBindTable to
>>    have unresolved dangling references to VPLS-A as long as VPLS-A is created before the PWBind
>>    Entry is set to active. On the other hand, it has been my experience that agent writers will not
>>    allow that mode of operation (i.e. will not allow creation of the dangling references) unless the MIB
>>    explicitly mandates it and I think such a mandate may be more trouble than it's worth. Therefore,
>>    to protect an NMS writer from ordering their create operations that way and not interoperating,
>>    I'd suggest explicitly saying that it is optional for the agent to allow entries to be created
>>    that point to non-existent VPLSConfig entries. The most straightforward path is for the NMS to create the
>>    VPLSConfig entry and then the vplsPWBind entries.
>>
>> The relevant tables are:
>>
>>       VplsPwBindEntry ::=
>>           SEQUENCE {
>>               vplsPwBindConfigType              INTEGER,
>>               vplsPwBindType                  INTEGER,
>>               vplsPwBindRowStatus             RowStatus,
>>               vplsPwBindStorageType             StorageType
>>           }
>>           INDEX  { vplsConfigIndex, pwIndex }
>>
>>       VplsBgpRteTargetEntry ::=
>>          SEQUENCE {
>>           vplsBgpRteTargetIndex          Unsigned32,
>>           vplsBgpRteTargetRTType         VplsBgpRouteTargetType,
>>           vplsBgpRteTargetRT             VplsBgpRouteTarget,
>>           vplsBgpRteTargetRowStatus      RowStatus
>>          }
>>          INDEX     { vplsConfigIndex, vplsBgpRteTargetIndex }
>>
>>      VplsBgpVEEntry ::= SEQUENCE {
>>           vplsBgpVEId          Unsigned32,
>>           vplsBgpVEName        SnmpAdminString,
>>           vplsBgpVEPreference  Unsigned32,
>>           vplsBgpVERowStatus   RowStatus,
>>           vplsBgpVEStorageType StorageType
>>         }
>>         INDEX  { vplsConfigIndex, vplsBgpVEId }
>>
>> 2. In a related matter, it would be good to define what happens to rows in the above three tables
>>    when the associated vplsConfigEntry is deleted. One option is nothing happens (i.e. allow the
>>    dangling references). Another option is language in the RowStatus objects like: "If this object
>>    is set to destroy(6) or deleted by the agent, all associated entries in the VplsPWBindTable,
>>    vplsBGPRteTargetTable and vplsBgpVETable shall be deleted."
>>
>>
>> 3. vplsConfigRowStatus
>>    vplsPwBindRowStatus
>>    vplsBgpADConfigRowStatus
>>    vplsBgpRteTargetRowStatus
>>    vplsBgpVERowStatus
>>
>>    What columns in those tables must be set to valid values before these RowStatus objects can be set
>>    to active? (Presumably all of them)? [Ref: RFC4181 4.6.4]
>>
>> 4. vplsPwBindEntry
>>    vplsBgpADConfigEntry
>>    vplsBgpRteTargetEntry
>>    vplsBgpVEEntry
>>
>>    As per RFC4181 4.6.4, "If the agent itself may also create and/or delete rows, then the conditions
>>    under which this can occur MUST be clearly documented in the row object DESCRIPTION clause."
>>
>>    Something like: "Entries in this table may be created or deleted through SNMP, as side-effects of
>>    console or other non-SNMP management commands, or upon learning via autodiscovery."
>>
>> 5. VplsBgpADConfigEntry
>>    VplsBgpRteTargetEntry
>>
>>    These tables need a StorageType object
>>
>> 6. vplsPwBindStorageType
>>    Presumably if this row was from auto-discovery, the storagetype for this row should be 'volatile' or
>>    'other'. My argument for 'other' is that the entry is 'algorithmically' stored.
>>    Further, you might want to document that if autodiscovered entries are deleted they would likely
>>    re-appear in the next autodiscovery interval. I'm guessing you don't intend to
>>    store a "negative" entry to keep the entry from reappearing. You'd have trouble deleting the
>>    negative entry anyway.
>>
>>    Similar issue for vplsBgpRteTargetRowStatus.
>>
>> 7. Page 6: Section 5: Example of the VPLS MIB modules usage
>>
>>    Bravo for including an example section!
>>
>> 8. This document is structured as three separate MIB modules. Did you consider putting everything in
>>    one MIB module and using the conformance section to modularize the MIB into 3 optional groups?
>>
>> 9. Page 15: vplsStatusTable
>>
>>   Do you intend to add a bunch of other stats to this table in the future? Since vplsStatusTable is
>>   a one-to-one mapping with vplsConfigTable and fate-shares with it, these could just be read-only
>>   variables in the vplsConfigTable and then this table would be unnecessary. . Further, if it is
>>   important that this be in a separate table then this should use the AUGMENTS clause, not
>>   INDEX. AUGMENTS is used for sub-tables that have a one-to-one mapping and that fate-share, like
>>   this one.
>>
>> 10. Page 23 vplsStatusChanged NOTIFICATION-TYPE
>>
>>   Need to specify which instances are included in the notification. Something like this: "The object
>>   instances included in the notification are the ones associated with the VPLS service
>>   whose status has changed."
>>
>>   vplsFwdFullAlarmRaised NOTIFICATION-TYPE
>>
>>   Similarly: "The object instances included in the notification are the ones associated with the
>>   VPLS service which has exceeded the threshold"
>>
>>   vplsFwdFullAlarmCleared NOTIFICATION-TYPE
>>
>>   Similarly: "The object instances included in the notification are the ones associated with the
>>   VPLS service which has fallen below the threshold."
>>
>> 11. Page 30: vplsLdpConfigMacAddrWithdraw: This object should be read-write, not read-create.
>>
>> 12. Generally: How does the agent know that you intend to start BGP or LDP on a VPLS entry that you
>> have created with SNMP?
>>
>> 13. Page 31: vplsLdpPwBindMacAddressLimit
>>
>>     Given there's no RowStatus I assume entries in the table are created by the agent as a side
>>     effect of the creation of PW bindings on an LDP-managed vpls service (regardless of whether the
>>     service or PW bindings were created manually, by signaling or
>>     by SNMP.
>>
>>     If this assumption is correct then the object should be read-write,
>>     not read-create.
>>
>> 14. Page 42: Security Considerations
>>
>>     In order to comply with rfc4181 Section 3.4, this should state explicitly that
>>     while the read-write and read-create objects must be protected by secure SNMP,
>>     none of them are especially disruptive. Similarly, while the read-only objects might present
>>     privacy concerns and due consideration should be given to protecting them with secure SNMP,
>>     none of these objects contain especially sensitive information.
>>
>> I also have a number of smaller items that I wrote notes about when I was on vacation. I need to collect those notes and
>> send them in another email.
>>
>>
>>
>> Steve
>>


--------------080809080104040308070905
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit

<html>
  <head>
    <meta content="text/html; charset=ISO-8859-1"
      http-equiv="Content-Type">
  </head>
  <body text="#000000" bgcolor="#FFFFFF">
    <div class="moz-cite-prefix">On 8/14/2013 2:07 PM, Agrahara Kiran
      Koushik wrote:<br>
    </div>
    <blockquote
      cite="mid:A689AE1D-9EBC-4160-A5C3-E4A653900E21@cisco.com"
      type="cite"><br>
      <blockquote type="cite">
        <pre wrap="">8. This document is structured as three separate MIB modules. Did you consider putting everything in
  one MIB module and using the conformance section to modularize the MIB into 3 optional groups?
</pre>
      </blockquote>
      <pre wrap="">
[kiran] We had specific requests in the beginning to separate the MIB into different modules instead
of one. We'd like to keep it that way if possible.</pre>
    </blockquote>
    It's not my preferred choice but it's a valid choice - I don't have
    a big problem with it. <br>
    <br>
    <blockquote
      cite="mid:A689AE1D-9EBC-4160-A5C3-E4A653900E21@cisco.com"
      type="cite"><br>
      <blockquote type="cite">
        <pre wrap="">12. Generally: How does the agent know that you intend to start BGP or LDP on a VPLS entry that you
have created with SNMP?
</pre>
      </blockquote>
      <pre wrap="">
[kiran] Not sure why it would be required?</pre>
    </blockquote>
    Here's what I was thinking:<br>
    Let's say a row is created via SNMP in the vplsConfigTable.
    Everything the agent knows about that VPLS service is contained in
    that row. I'm assuming the agent will need to be told whether to use
    BGP or LDP for that service.<br>
    <br>
    Perhaps it's a side-effect of whether you created an entry in the
    VplsLdpConfigTable or created one in the vplsBgpConfigTable? Could
    work I guess but it would be good to prohibit situations where
    neither entry is created or both are created (especially the
    latter). Most important would be to define authoritatively how the
    agent knows. It seemed like the other BDP and LDP tables could be
    empty even if the corresponding service was enabled so they wouldn't
    be good choices.<br>
    <br>
    Thanks,<br>
    Steve<br>
    <br>
    <br>
    <blockquote
      cite="mid:A689AE1D-9EBC-4160-A5C3-E4A653900E21@cisco.com"
      type="cite">
      <pre wrap="">


</pre>
      <br>
      <fieldset class="mimeAttachmentHeader"></fieldset>
      <br>
      <pre wrap="">

Thanks
Kiran


On Jul 16, 2013, at 8:02 PM, Steve Waldbusser <a class="moz-txt-link-rfc2396E" href="mailto:waldbusser@nextbeacon.com">&lt;waldbusser@nextbeacon.com&gt;</a> wrote:

</pre>
      <blockquote type="cite">
        <pre wrap="">Hello,

Here's my review of the the VPLS MIB.

1. There are three tables that have a many-to-one relationship with the vplsConfigTable AND that
  allow SNMP create operations. A potential ambiguity exists in the creation process for these
  tables. For example one could create entries in the vplsPwBindTable that reference VPLS entries
  that don't yet exist: IOW one could create vplsPWBindRowStatus.VPLS-A.PW-B and THEN create vpls-A
  in the vplsConfigTable. It is perfectly alright for the inactive rows in the vplsPWBindTable to
  have unresolved dangling references to VPLS-A as long as VPLS-A is created before the PWBind
  Entry is set to active. On the other hand, it has been my experience that agent writers will not
  allow that mode of operation (i.e. will not allow creation of the dangling references) unless the MIB
  explicitly mandates it and I think such a mandate may be more trouble than it's worth. Therefore,
  to protect an NMS writer from ordering their create operations that way and not interoperating,
  I'd suggest explicitly saying that it is optional for the agent to allow entries to be created
  that point to non-existent VPLSConfig entries. The most straightforward path is for the NMS to create the
  VPLSConfig entry and then the vplsPWBind entries.

The relevant tables are:

     VplsPwBindEntry ::=
         SEQUENCE {
             vplsPwBindConfigType              INTEGER,
             vplsPwBindType                  INTEGER,
             vplsPwBindRowStatus             RowStatus,
             vplsPwBindStorageType             StorageType
         }
         INDEX  { vplsConfigIndex, pwIndex }

     VplsBgpRteTargetEntry ::=
        SEQUENCE {
         vplsBgpRteTargetIndex          Unsigned32,
         vplsBgpRteTargetRTType         VplsBgpRouteTargetType,
         vplsBgpRteTargetRT             VplsBgpRouteTarget,
         vplsBgpRteTargetRowStatus      RowStatus
        }
        INDEX     { vplsConfigIndex, vplsBgpRteTargetIndex }

    VplsBgpVEEntry ::= SEQUENCE {
         vplsBgpVEId          Unsigned32,
         vplsBgpVEName        SnmpAdminString,
         vplsBgpVEPreference  Unsigned32,
         vplsBgpVERowStatus   RowStatus,
         vplsBgpVEStorageType StorageType
       }
       INDEX  { vplsConfigIndex, vplsBgpVEId }

2. In a related matter, it would be good to define what happens to rows in the above three tables
  when the associated vplsConfigEntry is deleted. One option is nothing happens (i.e. allow the
  dangling references). Another option is language in the RowStatus objects like: "If this object
  is set to destroy(6) or deleted by the agent, all associated entries in the VplsPWBindTable,
  vplsBGPRteTargetTable and vplsBgpVETable shall be deleted."


3. vplsConfigRowStatus
  vplsPwBindRowStatus
  vplsBgpADConfigRowStatus
  vplsBgpRteTargetRowStatus
  vplsBgpVERowStatus

  What columns in those tables must be set to valid values before these RowStatus objects can be set
  to active? (Presumably all of them)? [Ref: RFC4181 4.6.4]

4. vplsPwBindEntry
  vplsBgpADConfigEntry
  vplsBgpRteTargetEntry
  vplsBgpVEEntry

  As per RFC4181 4.6.4, "If the agent itself may also create and/or delete rows, then the conditions
  under which this can occur MUST be clearly documented in the row object DESCRIPTION clause."

  Something like: "Entries in this table may be created or deleted through SNMP, as side-effects of
  console or other non-SNMP management commands, or upon learning via autodiscovery."

5. VplsBgpADConfigEntry
  VplsBgpRteTargetEntry

  These tables need a StorageType object

6. vplsPwBindStorageType
  Presumably if this row was from auto-discovery, the storagetype for this row should be 'volatile' or
  'other'. My argument for 'other' is that the entry is 'algorithmically' stored.
  Further, you might want to document that if autodiscovered entries are deleted they would likely
  re-appear in the next autodiscovery interval. I'm guessing you don't intend to
  store a "negative" entry to keep the entry from reappearing. You'd have trouble deleting the
  negative entry anyway.

  Similar issue for vplsBgpRteTargetRowStatus.

7. Page 6: Section 5: Example of the VPLS MIB modules usage

  Bravo for including an example section!

8. This document is structured as three separate MIB modules. Did you consider putting everything in
  one MIB module and using the conformance section to modularize the MIB into 3 optional groups?

9. Page 15: vplsStatusTable

 Do you intend to add a bunch of other stats to this table in the future? Since vplsStatusTable is
 a one-to-one mapping with vplsConfigTable and fate-shares with it, these could just be read-only
 variables in the vplsConfigTable and then this table would be unnecessary. . Further, if it is
 important that this be in a separate table then this should use the AUGMENTS clause, not
 INDEX. AUGMENTS is used for sub-tables that have a one-to-one mapping and that fate-share, like
 this one.

10. Page 23 vplsStatusChanged NOTIFICATION-TYPE

 Need to specify which instances are included in the notification. Something like this: "The object
 instances included in the notification are the ones associated with the VPLS service
 whose status has changed."

 vplsFwdFullAlarmRaised NOTIFICATION-TYPE

 Similarly: "The object instances included in the notification are the ones associated with the
 VPLS service which has exceeded the threshold"

 vplsFwdFullAlarmCleared NOTIFICATION-TYPE

 Similarly: "The object instances included in the notification are the ones associated with the
 VPLS service which has fallen below the threshold."

11. Page 30: vplsLdpConfigMacAddrWithdraw: This object should be read-write, not read-create.

12. Generally: How does the agent know that you intend to start BGP or LDP on a VPLS entry that you
have created with SNMP?

13. Page 31: vplsLdpPwBindMacAddressLimit

   Given there's no RowStatus I assume entries in the table are created by the agent as a side
   effect of the creation of PW bindings on an LDP-managed vpls service (regardless of whether the
   service or PW bindings were created manually, by signaling or
   by SNMP.

   If this assumption is correct then the object should be read-write,
   not read-create.

14. Page 42: Security Considerations

   In order to comply with rfc4181 Section 3.4, this should state explicitly that
   while the read-write and read-create objects must be protected by secure SNMP,
   none of them are especially disruptive. Similarly, while the read-only objects might present
   privacy concerns and due consideration should be given to protecting them with secure SNMP,
   none of these objects contain especially sensitive information.

I also have a number of smaller items that I wrote notes about when I was on vacation. I need to collect those notes and
send them in another email.



Steve

</pre>
      </blockquote>
      <pre wrap="">
</pre>
    </blockquote>
    <br>
  </body>
</html>

--------------080809080104040308070905--

From waldbusser@nextbeacon.com  Wed Aug 14 22:13:19 2013
Return-Path: <waldbusser@nextbeacon.com>
X-Original-To: mib-doctors@ietfa.amsl.com
Delivered-To: mib-doctors@ietfa.amsl.com
Received: from localhost (localhost [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id B044511E8121 for <mib-doctors@ietfa.amsl.com>; Wed, 14 Aug 2013 22:13:19 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.598
X-Spam-Level: 
X-Spam-Status: No, score=-2.598 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, HTML_MESSAGE=0.001]
Received: from mail.ietf.org ([12.22.58.30]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id WLq5BaMsXygq for <mib-doctors@ietfa.amsl.com>; Wed, 14 Aug 2013 22:13:13 -0700 (PDT)
Received: from mxout-08.mxes.net (mxout-08.mxes.net [216.86.168.183]) by ietfa.amsl.com (Postfix) with ESMTP id 8765B11E8108 for <mib-doctors@ietf.org>; Wed, 14 Aug 2013 22:13:13 -0700 (PDT)
Received: from [127.0.0.1] (unknown [24.6.159.106]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by smtp.mxes.net (Postfix) with ESMTPSA id D97CA509B6; Thu, 15 Aug 2013 01:13:09 -0400 (EDT)
Message-ID: <520C6364.30107@nextbeacon.com>
Date: Wed, 14 Aug 2013 22:13:08 -0700
From: Steve Waldbusser <waldbusser@nextbeacon.com>
User-Agent: Mozilla/5.0 (Windows NT 5.1; rv:17.0) Gecko/20130801 Thunderbird/17.0.8
MIME-Version: 1.0
To: Agrahara Kiran Koushik <kkoushik@cisco.com>
References: <51E5ED0F.10207@nextbeacon.com> <A689AE1D-9EBC-4160-A5C3-E4A653900E21@cisco.com> <520C0FA7.1010904@nextbeacon.com> <90DE550C-5DB1-4A17-8606-84D04F5AFB44@cisco.com>
In-Reply-To: <90DE550C-5DB1-4A17-8606-84D04F5AFB44@cisco.com>
Content-Type: multipart/alternative; boundary="------------030106010003010804020301"
X-Antivirus: avast! (VPS 130814-1, 08/14/2013), Outbound message
X-Antivirus-Status: Clean
Cc: mib-doctors@ietf.org, "Giles Heron \(giheron\)" <giheron@cisco.com>, draft-ietf-l2vpn-vpls-mib.all@tools.ietf.org
Subject: Re: [MIB-DOCTORS] Review of draft-ietf-l2vpn-vpls-mib-10
X-BeenThere: mib-doctors@ietf.org
X-Mailman-Version: 2.1.12
Precedence: list
List-Id: MIB Doctors list <mib-doctors.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/mib-doctors>, <mailto:mib-doctors-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/mib-doctors>
List-Post: <mailto:mib-doctors@ietf.org>
List-Help: <mailto:mib-doctors-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/mib-doctors>, <mailto:mib-doctors-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 15 Aug 2013 05:13:19 -0000

This is a multi-part message in MIME format.
--------------030106010003010804020301
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

That's exactly what I was thinking (enumerated integer to be more 
exact). And it could be theoretically extended in the future if other 
signalling types are invented.

Steve

On 8/14/2013 4:35 PM, Agrahara Kiran Koushik wrote:
> Thanks for your input Steve.
>
> Do you think adding a new object in the vplsConfigTable to indicate 
> whether BGP or LDP is used for that service
> is sufficient to handle this ambiguity?
>
> Thanks
> Kiran
>
>
>>>> 12. Generally: How does the agent know that you intend to start BGP or LDP on a VPLS entry that you
>>>> have created with SNMP?
>>> [kiran] Not sure why it would be required?
>> Here's what I was thinking:
>> Let's say a row is created via SNMP in the vplsConfigTable. 
>> Everything the agent knows about that VPLS service is contained in 
>> that row. I'm assuming the agent will need to be told whether to use 
>> BGP or LDP for that service.
>>
>> Perhaps it's a side-effect of whether you created an entry in the 
>> VplsLdpConfigTable or created one in the vplsBgpConfigTable? Could 
>> work I guess but it would be good to prohibit situations where 
>> neither entry is created or both are created (especially the latter). 
>> Most important would be to define authoritatively how the agent 
>> knows. It seemed like the other BDP and LDP tables could be empty 
>> even if the corresponding service was enabled so they wouldn't be 
>> good choices.
>>
>> Thanks,
>> Steve
>>
>>
>>>
>>>
>>> Thanks
>>> Kiran
>>>
>>>
>>> On Jul 16, 2013, at 8:02 PM, Steve Waldbusser<waldbusser@nextbeacon.com>  wrote:
>>>
>>>> Hello,
>>>>
>>>> Here's my review of the the VPLS MIB.
>>>>
>>>> 1. There are three tables that have a many-to-one relationship with the vplsConfigTable AND that
>>>>    allow SNMP create operations. A potential ambiguity exists in the creation process for these
>>>>    tables. For example one could create entries in the vplsPwBindTable that reference VPLS entries
>>>>    that don't yet exist: IOW one could create vplsPWBindRowStatus.VPLS-A.PW-B and THEN create vpls-A
>>>>    in the vplsConfigTable. It is perfectly alright for the inactive rows in the vplsPWBindTable to
>>>>    have unresolved dangling references to VPLS-A as long as VPLS-A is created before the PWBind
>>>>    Entry is set to active. On the other hand, it has been my experience that agent writers will not
>>>>    allow that mode of operation (i.e. will not allow creation of the dangling references) unless the MIB
>>>>    explicitly mandates it and I think such a mandate may be more trouble than it's worth. Therefore,
>>>>    to protect an NMS writer from ordering their create operations that way and not interoperating,
>>>>    I'd suggest explicitly saying that it is optional for the agent to allow entries to be created
>>>>    that point to non-existent VPLSConfig entries. The most straightforward path is for the NMS to create the
>>>>    VPLSConfig entry and then the vplsPWBind entries.
>>>>
>>>> The relevant tables are:
>>>>
>>>>       VplsPwBindEntry ::=
>>>>           SEQUENCE {
>>>>               vplsPwBindConfigType              INTEGER,
>>>>               vplsPwBindType                  INTEGER,
>>>>               vplsPwBindRowStatus             RowStatus,
>>>>               vplsPwBindStorageType             StorageType
>>>>           }
>>>>           INDEX  { vplsConfigIndex, pwIndex }
>>>>
>>>>       VplsBgpRteTargetEntry ::=
>>>>          SEQUENCE {
>>>>           vplsBgpRteTargetIndex          Unsigned32,
>>>>           vplsBgpRteTargetRTType         VplsBgpRouteTargetType,
>>>>           vplsBgpRteTargetRT             VplsBgpRouteTarget,
>>>>           vplsBgpRteTargetRowStatus      RowStatus
>>>>          }
>>>>          INDEX     { vplsConfigIndex, vplsBgpRteTargetIndex }
>>>>
>>>>      VplsBgpVEEntry ::= SEQUENCE {
>>>>           vplsBgpVEId          Unsigned32,
>>>>           vplsBgpVEName        SnmpAdminString,
>>>>           vplsBgpVEPreference  Unsigned32,
>>>>           vplsBgpVERowStatus   RowStatus,
>>>>           vplsBgpVEStorageType StorageType
>>>>         }
>>>>         INDEX  { vplsConfigIndex, vplsBgpVEId }
>>>>
>>>> 2. In a related matter, it would be good to define what happens to rows in the above three tables
>>>>    when the associated vplsConfigEntry is deleted. One option is nothing happens (i.e. allow the
>>>>    dangling references). Another option is language in the RowStatus objects like: "If this object
>>>>    is set to destroy(6) or deleted by the agent, all associated entries in the VplsPWBindTable,
>>>>    vplsBGPRteTargetTable and vplsBgpVETable shall be deleted."
>>>>
>>>>
>>>> 3. vplsConfigRowStatus
>>>>    vplsPwBindRowStatus
>>>>    vplsBgpADConfigRowStatus
>>>>    vplsBgpRteTargetRowStatus
>>>>    vplsBgpVERowStatus
>>>>
>>>>    What columns in those tables must be set to valid values before these RowStatus objects can be set
>>>>    to active? (Presumably all of them)? [Ref: RFC4181 4.6.4]
>>>>
>>>> 4. vplsPwBindEntry
>>>>    vplsBgpADConfigEntry
>>>>    vplsBgpRteTargetEntry
>>>>    vplsBgpVEEntry
>>>>
>>>>    As per RFC4181 4.6.4, "If the agent itself may also create and/or delete rows, then the conditions
>>>>    under which this can occur MUST be clearly documented in the row object DESCRIPTION clause."
>>>>
>>>>    Something like: "Entries in this table may be created or deleted through SNMP, as side-effects of
>>>>    console or other non-SNMP management commands, or upon learning via autodiscovery."
>>>>
>>>> 5. VplsBgpADConfigEntry
>>>>    VplsBgpRteTargetEntry
>>>>
>>>>    These tables need a StorageType object
>>>>
>>>> 6. vplsPwBindStorageType
>>>>    Presumably if this row was from auto-discovery, the storagetype for this row should be 'volatile' or
>>>>    'other'. My argument for 'other' is that the entry is 'algorithmically' stored.
>>>>    Further, you might want to document that if autodiscovered entries are deleted they would likely
>>>>    re-appear in the next autodiscovery interval. I'm guessing you don't intend to
>>>>    store a "negative" entry to keep the entry from reappearing. You'd have trouble deleting the
>>>>    negative entry anyway.
>>>>
>>>>    Similar issue for vplsBgpRteTargetRowStatus.
>>>>
>>>> 7. Page 6: Section 5: Example of the VPLS MIB modules usage
>>>>
>>>>    Bravo for including an example section!
>>>>
>>>> 8. This document is structured as three separate MIB modules. Did you consider putting everything in
>>>>    one MIB module and using the conformance section to modularize the MIB into 3 optional groups?
>>>>
>>>> 9. Page 15: vplsStatusTable
>>>>
>>>>   Do you intend to add a bunch of other stats to this table in the future? Since vplsStatusTable is
>>>>   a one-to-one mapping with vplsConfigTable and fate-shares with it, these could just be read-only
>>>>   variables in the vplsConfigTable and then this table would be unnecessary. . Further, if it is
>>>>   important that this be in a separate table then this should use the AUGMENTS clause, not
>>>>   INDEX. AUGMENTS is used for sub-tables that have a one-to-one mapping and that fate-share, like
>>>>   this one.
>>>>
>>>> 10. Page 23 vplsStatusChanged NOTIFICATION-TYPE
>>>>
>>>>   Need to specify which instances are included in the notification. Something like this: "The object
>>>>   instances included in the notification are the ones associated with the VPLS service
>>>>   whose status has changed."
>>>>
>>>>   vplsFwdFullAlarmRaised NOTIFICATION-TYPE
>>>>
>>>>   Similarly: "The object instances included in the notification are the ones associated with the
>>>>   VPLS service which has exceeded the threshold"
>>>>
>>>>   vplsFwdFullAlarmCleared NOTIFICATION-TYPE
>>>>
>>>>   Similarly: "The object instances included in the notification are the ones associated with the
>>>>   VPLS service which has fallen below the threshold."
>>>>
>>>> 11. Page 30: vplsLdpConfigMacAddrWithdraw: This object should be read-write, not read-create.
>>>>
>>>> 12. Generally: How does the agent know that you intend to start BGP or LDP on a VPLS entry that you
>>>> have created with SNMP?
>>>>
>>>> 13. Page 31: vplsLdpPwBindMacAddressLimit
>>>>
>>>>     Given there's no RowStatus I assume entries in the table are created by the agent as a side
>>>>     effect of the creation of PW bindings on an LDP-managed vpls service (regardless of whether the
>>>>     service or PW bindings were created manually, by signaling or
>>>>     by SNMP.
>>>>
>>>>     If this assumption is correct then the object should be read-write,
>>>>     not read-create.
>>>>
>>>> 14. Page 42: Security Considerations
>>>>
>>>>     In order to comply with rfc4181 Section 3.4, this should state explicitly that
>>>>     while the read-write and read-create objects must be protected by secure SNMP,
>>>>     none of them are especially disruptive. Similarly, while the read-only objects might present
>>>>     privacy concerns and due consideration should be given to protecting them with secure SNMP,
>>>>     none of these objects contain especially sensitive information.
>>>>
>>>> I also have a number of smaller items that I wrote notes about when I was on vacation. I need to collect those notes and
>>>> send them in another email.
>>>>
>>>>
>>>>
>>>> Steve
>>>>
>>
>


--------------030106010003010804020301
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit

<html>
  <head>
    <meta content="text/html; charset=ISO-8859-1"
      http-equiv="Content-Type">
  </head>
  <body text="#000000" bgcolor="#FFFFFF">
    <div class="moz-cite-prefix">That's exactly what I was thinking
      (enumerated integer to be more exact). And it could be
      theoretically extended in the future if other signalling types are
      invented.<br>
      <br>
      Steve<br>
      <br>
      On 8/14/2013 4:35 PM, Agrahara Kiran Koushik wrote:<br>
    </div>
    <blockquote
      cite="mid:90DE550C-5DB1-4A17-8606-84D04F5AFB44@cisco.com"
      type="cite">
      <meta http-equiv="Content-Type" content="text/html;
        charset=ISO-8859-1">
      Thanks for your input Steve.
      <div><br>
      </div>
      <div>Do you think adding a new object in the vplsConfigTable to
        indicate whether BGP or LDP is used for that service</div>
      <div>is sufficient to handle this ambiguity?</div>
      <div><br>
      </div>
      <div>Thanks</div>
      <div>Kiran</div>
      <div><br>
        <div>
          <div><br>
            <blockquote type="cite">
              <div text="#000000" bgcolor="#FFFFFF">
                <blockquote
                  cite="mid:A689AE1D-9EBC-4160-A5C3-E4A653900E21@cisco.com"
                  type="cite">
                  <blockquote type="cite">
                    <pre wrap="">12. Generally: How does the agent know that you intend to start BGP or LDP on a VPLS entry that you
have created with SNMP?
</pre>
                  </blockquote>
                  <pre wrap="">[kiran] Not sure why it would be required?</pre>
                </blockquote>
                Here's what I was thinking:<br>
                Let's say a row is created via SNMP in the
                vplsConfigTable. Everything the agent knows about that
                VPLS service is contained in that row. I'm assuming the
                agent will need to be told whether to use BGP or LDP for
                that service.<br>
                <br>
                Perhaps it's a side-effect of whether you created an
                entry in the VplsLdpConfigTable or created one in the
                vplsBgpConfigTable? Could work I guess but it would be
                good to prohibit situations where neither entry is
                created or both are created (especially the latter).
                Most important would be to define authoritatively how
                the agent knows. It seemed like the other BDP and LDP
                tables could be empty even if the corresponding service
                was enabled so they wouldn't be good choices.<br>
                <br>
                Thanks,<br>
                Steve<br>
                <br>
                <br>
                <blockquote
                  cite="mid:A689AE1D-9EBC-4160-A5C3-E4A653900E21@cisco.com"
                  type="cite">
                  <pre wrap="">
</pre>
                  <br>
                  <fieldset class="mimeAttachmentHeader"></fieldset>
                  <br>
                  <pre wrap="">Thanks
Kiran


On Jul 16, 2013, at 8:02 PM, Steve Waldbusser <a moz-do-not-send="true" class="moz-txt-link-rfc2396E" href="mailto:waldbusser@nextbeacon.com">&lt;waldbusser@nextbeacon.com&gt;</a> wrote:

</pre>
                  <blockquote type="cite">
                    <pre wrap="">Hello,

Here's my review of the the VPLS MIB.

1. There are three tables that have a many-to-one relationship with the vplsConfigTable AND that
  allow SNMP create operations. A potential ambiguity exists in the creation process for these
  tables. For example one could create entries in the vplsPwBindTable that reference VPLS entries
  that don't yet exist: IOW one could create vplsPWBindRowStatus.VPLS-A.PW-B and THEN create vpls-A
  in the vplsConfigTable. It is perfectly alright for the inactive rows in the vplsPWBindTable to
  have unresolved dangling references to VPLS-A as long as VPLS-A is created before the PWBind
  Entry is set to active. On the other hand, it has been my experience that agent writers will not
  allow that mode of operation (i.e. will not allow creation of the dangling references) unless the MIB
  explicitly mandates it and I think such a mandate may be more trouble than it's worth. Therefore,
  to protect an NMS writer from ordering their create operations that way and not interoperating,
  I'd suggest explicitly saying that it is optional for the agent to allow entries to be created
  that point to non-existent VPLSConfig entries. The most straightforward path is for the NMS to create the
  VPLSConfig entry and then the vplsPWBind entries.

The relevant tables are:

     VplsPwBindEntry ::=
         SEQUENCE {
             vplsPwBindConfigType              INTEGER,
             vplsPwBindType                  INTEGER,
             vplsPwBindRowStatus             RowStatus,
             vplsPwBindStorageType             StorageType
         }
         INDEX  { vplsConfigIndex, pwIndex }

     VplsBgpRteTargetEntry ::=
        SEQUENCE {
         vplsBgpRteTargetIndex          Unsigned32,
         vplsBgpRteTargetRTType         VplsBgpRouteTargetType,
         vplsBgpRteTargetRT             VplsBgpRouteTarget,
         vplsBgpRteTargetRowStatus      RowStatus
        }
        INDEX     { vplsConfigIndex, vplsBgpRteTargetIndex }

    VplsBgpVEEntry ::= SEQUENCE {
         vplsBgpVEId          Unsigned32,
         vplsBgpVEName        SnmpAdminString,
         vplsBgpVEPreference  Unsigned32,
         vplsBgpVERowStatus   RowStatus,
         vplsBgpVEStorageType StorageType
       }
       INDEX  { vplsConfigIndex, vplsBgpVEId }

2. In a related matter, it would be good to define what happens to rows in the above three tables
  when the associated vplsConfigEntry is deleted. One option is nothing happens (i.e. allow the
  dangling references). Another option is language in the RowStatus objects like: "If this object
  is set to destroy(6) or deleted by the agent, all associated entries in the VplsPWBindTable,
  vplsBGPRteTargetTable and vplsBgpVETable shall be deleted."


3. vplsConfigRowStatus
  vplsPwBindRowStatus
  vplsBgpADConfigRowStatus
  vplsBgpRteTargetRowStatus
  vplsBgpVERowStatus

  What columns in those tables must be set to valid values before these RowStatus objects can be set
  to active? (Presumably all of them)? [Ref: RFC4181 4.6.4]

4. vplsPwBindEntry
  vplsBgpADConfigEntry
  vplsBgpRteTargetEntry
  vplsBgpVEEntry

  As per RFC4181 4.6.4, "If the agent itself may also create and/or delete rows, then the conditions
  under which this can occur MUST be clearly documented in the row object DESCRIPTION clause."

  Something like: "Entries in this table may be created or deleted through SNMP, as side-effects of
  console or other non-SNMP management commands, or upon learning via autodiscovery."

5. VplsBgpADConfigEntry
  VplsBgpRteTargetEntry

  These tables need a StorageType object

6. vplsPwBindStorageType
  Presumably if this row was from auto-discovery, the storagetype for this row should be 'volatile' or
  'other'. My argument for 'other' is that the entry is 'algorithmically' stored.
  Further, you might want to document that if autodiscovered entries are deleted they would likely
  re-appear in the next autodiscovery interval. I'm guessing you don't intend to
  store a "negative" entry to keep the entry from reappearing. You'd have trouble deleting the
  negative entry anyway.

  Similar issue for vplsBgpRteTargetRowStatus.

7. Page 6: Section 5: Example of the VPLS MIB modules usage

  Bravo for including an example section!

8. This document is structured as three separate MIB modules. Did you consider putting everything in
  one MIB module and using the conformance section to modularize the MIB into 3 optional groups?

9. Page 15: vplsStatusTable

 Do you intend to add a bunch of other stats to this table in the future? Since vplsStatusTable is
 a one-to-one mapping with vplsConfigTable and fate-shares with it, these could just be read-only
 variables in the vplsConfigTable and then this table would be unnecessary. . Further, if it is
 important that this be in a separate table then this should use the AUGMENTS clause, not
 INDEX. AUGMENTS is used for sub-tables that have a one-to-one mapping and that fate-share, like
 this one.

10. Page 23 vplsStatusChanged NOTIFICATION-TYPE

 Need to specify which instances are included in the notification. Something like this: "The object
 instances included in the notification are the ones associated with the VPLS service
 whose status has changed."

 vplsFwdFullAlarmRaised NOTIFICATION-TYPE

 Similarly: "The object instances included in the notification are the ones associated with the
 VPLS service which has exceeded the threshold"

 vplsFwdFullAlarmCleared NOTIFICATION-TYPE

 Similarly: "The object instances included in the notification are the ones associated with the
 VPLS service which has fallen below the threshold."

11. Page 30: vplsLdpConfigMacAddrWithdraw: This object should be read-write, not read-create.

12. Generally: How does the agent know that you intend to start BGP or LDP on a VPLS entry that you
have created with SNMP?

13. Page 31: vplsLdpPwBindMacAddressLimit

   Given there's no RowStatus I assume entries in the table are created by the agent as a side
   effect of the creation of PW bindings on an LDP-managed vpls service (regardless of whether the
   service or PW bindings were created manually, by signaling or
   by SNMP.

   If this assumption is correct then the object should be read-write,
   not read-create.

14. Page 42: Security Considerations

   In order to comply with rfc4181 Section 3.4, this should state explicitly that
   while the read-write and read-create objects must be protected by secure SNMP,
   none of them are especially disruptive. Similarly, while the read-only objects might present
   privacy concerns and due consideration should be given to protecting them with secure SNMP,
   none of these objects contain especially sensitive information.

I also have a number of smaller items that I wrote notes about when I was on vacation. I need to collect those notes and
send them in another email.



Steve

</pre>
                  </blockquote>
                </blockquote>
                <br>
              </div>
            </blockquote>
          </div>
          <br>
        </div>
      </div>
    </blockquote>
    <br>
  </body>
</html>

--------------030106010003010804020301--

From giheron@cisco.com  Thu Aug 15 01:23:43 2013
Return-Path: <giheron@cisco.com>
X-Original-To: mib-doctors@ietfa.amsl.com
Delivered-To: mib-doctors@ietfa.amsl.com
Received: from localhost (localhost [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 4075111E80C5 for <mib-doctors@ietfa.amsl.com>; Thu, 15 Aug 2013 01:23:38 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -10.599
X-Spam-Level: 
X-Spam-Status: No, score=-10.599 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, RCVD_IN_DNSWL_HI=-8]
Received: from mail.ietf.org ([12.22.58.30]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id DUCPmPGCuufs for <mib-doctors@ietfa.amsl.com>; Thu, 15 Aug 2013 01:23:31 -0700 (PDT)
Received: from rcdn-iport-1.cisco.com (rcdn-iport-1.cisco.com [173.37.86.72]) by ietfa.amsl.com (Postfix) with ESMTP id 1FBFA11E80D5 for <mib-doctors@ietf.org>; Thu, 15 Aug 2013 01:23:27 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=@cisco.com; l=10202; q=dns/txt; s=iport; t=1376555007; x=1377764607; h=from:to:cc:subject:date:message-id:references: in-reply-to:content-id:content-transfer-encoding: mime-version; bh=zcVJ7sh/Lo1tMbiBT0ACl7TEhWCrcMmqnmf3qktK9cU=; b=GYTT58SfMzRODjIfyS+JX+W28B4qTT52bUWNQcoQI37nUC+g/drEyVW9 hsJh4qP4MM4ZpwuNROjKRxQJ5TGcIJGzmqynSbeLItKMQCW2fY+HBYzDO mtUej7dQaUr36jIr/1O4QxQq0lzhBO2i6gVnVYvMOCN8vohSzGF4xhfHV A=;
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: AjwFAMiODFKtJV2d/2dsb2JhbABbgwaBBb8PgSAWdIIkAQEBAwFtBQcFCwIBCBgKJDIlAgQOBQiIAga6AY8DAYEZAjEHgxt3A4h1ixiVKYMbgWkIFyI
X-IronPort-AV: E=Sophos;i="4.89,883,1367971200"; d="scan'208";a="247403031"
Received: from rcdn-core-6.cisco.com ([173.37.93.157]) by rcdn-iport-1.cisco.com with ESMTP; 15 Aug 2013 08:23:26 +0000
Received: from xhc-aln-x07.cisco.com (xhc-aln-x07.cisco.com [173.36.12.81]) by rcdn-core-6.cisco.com (8.14.5/8.14.5) with ESMTP id r7F8NQwq012947 (version=TLSv1/SSLv3 cipher=AES128-SHA bits=128 verify=FAIL); Thu, 15 Aug 2013 08:23:26 GMT
Received: from xmb-rcd-x04.cisco.com ([169.254.8.180]) by xhc-aln-x07.cisco.com ([173.36.12.81]) with mapi id 14.02.0318.004; Thu, 15 Aug 2013 03:23:25 -0500
From: "Giles Heron (giheron)" <giheron@cisco.com>
To: Steve Waldbusser <waldbusser@nextbeacon.com>
Thread-Topic: Review of draft-ietf-l2vpn-vpls-mib-10
Thread-Index: AQHOgolKKhIVtXkP0UGrio1HBfqJW5mVszIAgAAj/ICAAAVqAIAAXmkAgAA1KwA=
Date: Thu, 15 Aug 2013 08:23:25 +0000
Message-ID: <136D0A04DAA0A64881D68D8374E6E0770911C2A4@xmb-rcd-x04.cisco.com>
References: <51E5ED0F.10207@nextbeacon.com> <A689AE1D-9EBC-4160-A5C3-E4A653900E21@cisco.com> <520C0FA7.1010904@nextbeacon.com> <90DE550C-5DB1-4A17-8606-84D04F5AFB44@cisco.com> <520C6364.30107@nextbeacon.com>
In-Reply-To: <520C6364.30107@nextbeacon.com>
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
x-originating-ip: [10.61.171.158]
Content-Type: text/plain; charset="iso-8859-1"
Content-ID: <EE508845FF2FCD45A8682CBCA80B96EE@emea.cisco.com>
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
Cc: "Kiran Koushik Agrahara Sreenivasa \(kkoushik\)" <kkoushik@cisco.com>, "<draft-ietf-l2vpn-vpls-mib.all@tools.ietf.org>" <draft-ietf-l2vpn-vpls-mib.all@tools.ietf.org>, "<mib-doctors@ietf.org>" <mib-doctors@ietf.org>
Subject: Re: [MIB-DOCTORS] Review of draft-ietf-l2vpn-vpls-mib-10
X-BeenThere: mib-doctors@ietf.org
X-Mailman-Version: 2.1.12
Precedence: list
List-Id: MIB Doctors list <mib-doctors.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/mib-doctors>, <mailto:mib-doctors-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/mib-doctors>
List-Post: <mailto:mib-doctors@ietf.org>
List-Help: <mailto:mib-doctors-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/mib-doctors>, <mailto:mib-doctors-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 15 Aug 2013 08:23:45 -0000

sounds sensible.

There's also a valid option of having static configuration of the PW labels=
 rather than using LDP or BGP signalling, so the enumerated integer probabl=
y needs values for "None", "LDP" and "BGP".

Giles

On 15 Aug 2013, at 06:13, Steve Waldbusser <waldbusser@nextbeacon.com> wrot=
e:

> That's exactly what I was thinking (enumerated integer to be more exact).=
 And it could be theoretically extended in the future if other signalling t=
ypes are invented.
>=20
> Steve
>=20
> On 8/14/2013 4:35 PM, Agrahara Kiran Koushik wrote:
>> Thanks for your input Steve.
>>=20
>> Do you think adding a new object in the vplsConfigTable to indicate whet=
her BGP or LDP is used for that service
>> is sufficient to handle this ambiguity?
>>=20
>> Thanks
>> Kiran
>>=20
>>=20
>>>>> 12. Generally: How does the agent know that you intend to start BGP o=
r LDP on a VPLS entry that you
>>>>> have created with SNMP?
>>>>>=20
>>>> [kiran] Not sure why it would be required?
>>> Here's what I was thinking:
>>> Let's say a row is created via SNMP in the vplsConfigTable. Everything =
the agent knows about that VPLS service is contained in that row. I'm assum=
ing the agent will need to be told whether to use BGP or LDP for that servi=
ce.
>>>=20
>>> Perhaps it's a side-effect of whether you created an entry in the VplsL=
dpConfigTable or created one in the vplsBgpConfigTable? Could work I guess =
but it would be good to prohibit situations where neither entry is created =
or both are created (especially the latter). Most important would be to def=
ine authoritatively how the agent knows. It seemed like the other BDP and L=
DP tables could be empty even if the corresponding service was enabled so t=
hey wouldn't be good choices.
>>>=20
>>> Thanks,
>>> Steve
>>>=20
>>>=20
>>>>=20
>>>>=20
>>>> Thanks
>>>> Kiran
>>>>=20
>>>>=20
>>>> On Jul 16, 2013, at 8:02 PM, Steve Waldbusser=20
>>>> <waldbusser@nextbeacon.com>
>>>>  wrote:
>>>>=20
>>>>=20
>>>>> Hello,
>>>>>=20
>>>>> Here's my review of the the VPLS MIB.
>>>>>=20
>>>>> 1. There are three tables that have a many-to-one relationship with t=
he vplsConfigTable AND that
>>>>>   allow SNMP create operations. A potential ambiguity exists in the c=
reation process for these
>>>>>   tables. For example one could create entries in the vplsPwBindTable=
 that reference VPLS entries
>>>>>   that don't yet exist: IOW one could create vplsPWBindRowStatus.VPLS=
-A.PW-B and THEN create vpls-A
>>>>>   in the vplsConfigTable. It is perfectly alright for the inactive ro=
ws in the vplsPWBindTable to
>>>>>   have unresolved dangling references to VPLS-A as long as VPLS-A is =
created before the PWBind
>>>>>   Entry is set to active. On the other hand, it has been my experienc=
e that agent writers will not
>>>>>   allow that mode of operation (i.e. will not allow creation of the d=
angling references) unless the MIB
>>>>>   explicitly mandates it and I think such a mandate may be more troub=
le than it's worth. Therefore,
>>>>>   to protect an NMS writer from ordering their create operations that=
 way and not interoperating,
>>>>>   I'd suggest explicitly saying that it is optional for the agent to =
allow entries to be created
>>>>>   that point to non-existent VPLSConfig entries. The most straightfor=
ward path is for the NMS to create the
>>>>>   VPLSConfig entry and then the vplsPWBind entries.
>>>>>=20
>>>>> The relevant tables are:
>>>>>=20
>>>>>      VplsPwBindEntry ::=3D
>>>>>          SEQUENCE {
>>>>>              vplsPwBindConfigType              INTEGER,
>>>>>              vplsPwBindType                  INTEGER,
>>>>>              vplsPwBindRowStatus             RowStatus,
>>>>>              vplsPwBindStorageType             StorageType
>>>>>          }
>>>>>          INDEX  { vplsConfigIndex, pwIndex }
>>>>>=20
>>>>>      VplsBgpRteTargetEntry ::=3D
>>>>>         SEQUENCE {
>>>>>          vplsBgpRteTargetIndex          Unsigned32,
>>>>>          vplsBgpRteTargetRTType         VplsBgpRouteTargetType,
>>>>>          vplsBgpRteTargetRT             VplsBgpRouteTarget,
>>>>>          vplsBgpRteTargetRowStatus      RowStatus
>>>>>         }
>>>>>         INDEX     { vplsConfigIndex, vplsBgpRteTargetIndex }
>>>>>=20
>>>>>     VplsBgpVEEntry ::=3D SEQUENCE {
>>>>>          vplsBgpVEId          Unsigned32,
>>>>>          vplsBgpVEName        SnmpAdminString,
>>>>>          vplsBgpVEPreference  Unsigned32,
>>>>>          vplsBgpVERowStatus   RowStatus,
>>>>>          vplsBgpVEStorageType StorageType
>>>>>        }
>>>>>        INDEX  { vplsConfigIndex, vplsBgpVEId }
>>>>>=20
>>>>> 2. In a related matter, it would be good to define what happens to ro=
ws in the above three tables
>>>>>   when the associated vplsConfigEntry is deleted. One option is nothi=
ng happens (i.e. allow the
>>>>>   dangling references). Another option is language in the RowStatus o=
bjects like: "If this object
>>>>>   is set to destroy(6) or deleted by the agent, all associated entrie=
s in the VplsPWBindTable,
>>>>>   vplsBGPRteTargetTable and vplsBgpVETable shall be deleted."
>>>>>=20
>>>>>=20
>>>>> 3. vplsConfigRowStatus
>>>>>   vplsPwBindRowStatus
>>>>>   vplsBgpADConfigRowStatus
>>>>>   vplsBgpRteTargetRowStatus
>>>>>   vplsBgpVERowStatus
>>>>>=20
>>>>>   What columns in those tables must be set to valid values before the=
se RowStatus objects can be set
>>>>>   to active? (Presumably all of them)? [Ref: RFC4181 4.6.4]
>>>>>=20
>>>>> 4. vplsPwBindEntry
>>>>>   vplsBgpADConfigEntry
>>>>>   vplsBgpRteTargetEntry
>>>>>   vplsBgpVEEntry
>>>>>=20
>>>>>   As per RFC4181 4.6.4, "If the agent itself may also create and/or d=
elete rows, then the conditions
>>>>>   under which this can occur MUST be clearly documented in the row ob=
ject DESCRIPTION clause."
>>>>>=20
>>>>>   Something like: "Entries in this table may be created or deleted th=
rough SNMP, as side-effects of
>>>>>   console or other non-SNMP management commands, or upon learning via=
 autodiscovery."
>>>>>=20
>>>>> 5. VplsBgpADConfigEntry
>>>>>   VplsBgpRteTargetEntry
>>>>>=20
>>>>>   These tables need a StorageType object
>>>>>=20
>>>>> 6. vplsPwBindStorageType
>>>>>   Presumably if this row was from auto-discovery, the storagetype for=
 this row should be 'volatile' or
>>>>>   'other'. My argument for 'other' is that the entry is 'algorithmica=
lly' stored.
>>>>>   Further, you might want to document that if autodiscovered entries =
are deleted they would likely
>>>>>   re-appear in the next autodiscovery interval. I'm guessing you don'=
t intend to
>>>>>   store a "negative" entry to keep the entry from reappearing. You'd =
have trouble deleting the
>>>>>   negative entry anyway.
>>>>>=20
>>>>>   Similar issue for vplsBgpRteTargetRowStatus.
>>>>>=20
>>>>> 7. Page 6: Section 5: Example of the VPLS MIB modules usage
>>>>>=20
>>>>>   Bravo for including an example section!
>>>>>=20
>>>>> 8. This document is structured as three separate MIB modules. Did you=
 consider putting everything in
>>>>>   one MIB module and using the conformance section to modularize the =
MIB into 3 optional groups?
>>>>>=20
>>>>> 9. Page 15: vplsStatusTable
>>>>>=20
>>>>>  Do you intend to add a bunch of other stats to this table in the fut=
ure? Since vplsStatusTable is
>>>>>  a one-to-one mapping with vplsConfigTable and fate-shares with it, t=
hese could just be read-only
>>>>>  variables in the vplsConfigTable and then this table would be unnece=
ssary. . Further, if it is
>>>>>  important that this be in a separate table then this should use the =
AUGMENTS clause, not
>>>>>  INDEX. AUGMENTS is used for sub-tables that have a one-to-one mappin=
g and that fate-share, like
>>>>>  this one.
>>>>>=20
>>>>> 10. Page 23 vplsStatusChanged NOTIFICATION-TYPE
>>>>>=20
>>>>>  Need to specify which instances are included in the notification. So=
mething like this: "The object
>>>>>  instances included in the notification are the ones associated with =
the VPLS service
>>>>>  whose status has changed."
>>>>>=20
>>>>>  vplsFwdFullAlarmRaised NOTIFICATION-TYPE
>>>>>=20
>>>>>  Similarly: "The object instances included in the notification are th=
e ones associated with the
>>>>>  VPLS service which has exceeded the threshold"
>>>>>=20
>>>>>  vplsFwdFullAlarmCleared NOTIFICATION-TYPE
>>>>>=20
>>>>>  Similarly: "The object instances included in the notification are th=
e ones associated with the
>>>>>  VPLS service which has fallen below the threshold."
>>>>>=20
>>>>> 11. Page 30: vplsLdpConfigMacAddrWithdraw: This object should be read=
-write, not read-create.
>>>>>=20
>>>>> 12. Generally: How does the agent know that you intend to start BGP o=
r LDP on a VPLS entry that you
>>>>> have created with SNMP?
>>>>>=20
>>>>> 13. Page 31: vplsLdpPwBindMacAddressLimit
>>>>>=20
>>>>>    Given there's no RowStatus I assume entries in the table are creat=
ed by the agent as a side
>>>>>    effect of the creation of PW bindings on an LDP-managed vpls servi=
ce (regardless of whether the
>>>>>    service or PW bindings were created manually, by signaling or
>>>>>    by SNMP.
>>>>>=20
>>>>>    If this assumption is correct then the object should be read-write=
,
>>>>>    not read-create.
>>>>>=20
>>>>> 14. Page 42: Security Considerations
>>>>>=20
>>>>>    In order to comply with rfc4181 Section 3.4, this should state exp=
licitly that
>>>>>    while the read-write and read-create objects must be protected by =
secure SNMP,
>>>>>    none of them are especially disruptive. Similarly, while the read-=
only objects might present
>>>>>    privacy concerns and due consideration should be given to protecti=
ng them with secure SNMP,
>>>>>    none of these objects contain especially sensitive information.
>>>>>=20
>>>>> I also have a number of smaller items that I wrote notes about when I=
 was on vacation. I need to collect those notes and
>>>>> send them in another email.
>>>>>=20
>>>>>=20
>>>>>=20
>>>>> Steve
>>>>>=20
>>>>>=20
>>>=20
>>=20
>=20


From waldbusser@nextbeacon.com  Thu Aug 15 13:36:59 2013
Return-Path: <waldbusser@nextbeacon.com>
X-Original-To: mib-doctors@ietfa.amsl.com
Delivered-To: mib-doctors@ietfa.amsl.com
Received: from localhost (localhost [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 7157A11E81D4 for <mib-doctors@ietfa.amsl.com>; Thu, 15 Aug 2013 13:36:59 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.599
X-Spam-Level: 
X-Spam-Status: No, score=-2.599 tagged_above=-999 required=5 tests=[AWL=0.001,  BAYES_00=-2.599]
Received: from mail.ietf.org ([12.22.58.30]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id w2YbsIX2YWVy for <mib-doctors@ietfa.amsl.com>; Thu, 15 Aug 2013 13:36:53 -0700 (PDT)
Received: from mxout-08.mxes.net (mxout-08.mxes.net [216.86.168.183]) by ietfa.amsl.com (Postfix) with ESMTP id 5D35611E8199 for <mib-doctors@ietf.org>; Thu, 15 Aug 2013 13:36:53 -0700 (PDT)
Received: from [127.0.0.1] (unknown [24.6.159.106]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by smtp.mxes.net (Postfix) with ESMTPSA id E270F50A85; Thu, 15 Aug 2013 16:36:49 -0400 (EDT)
Message-ID: <520D3BE0.9010803@nextbeacon.com>
Date: Thu, 15 Aug 2013 13:36:48 -0700
From: Steve Waldbusser <waldbusser@nextbeacon.com>
User-Agent: Mozilla/5.0 (Windows NT 5.1; rv:17.0) Gecko/20130801 Thunderbird/17.0.8
MIME-Version: 1.0
To: "Giles Heron (giheron)" <giheron@cisco.com>
References: <51E5ED0F.10207@nextbeacon.com> <A689AE1D-9EBC-4160-A5C3-E4A653900E21@cisco.com> <520C0FA7.1010904@nextbeacon.com> <90DE550C-5DB1-4A17-8606-84D04F5AFB44@cisco.com> <520C6364.30107@nextbeacon.com> <136D0A04DAA0A64881D68D8374E6E0770911C2A4@xmb-rcd-x04.cisco.com>
In-Reply-To: <136D0A04DAA0A64881D68D8374E6E0770911C2A4@xmb-rcd-x04.cisco.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-Antivirus: avast! (VPS 130815-1, 08/15/2013), Outbound message
X-Antivirus-Status: Clean
Cc: "Kiran Koushik Agrahara Sreenivasa \(kkoushik\)" <kkoushik@cisco.com>, "<draft-ietf-l2vpn-vpls-mib.all@tools.ietf.org>" <draft-ietf-l2vpn-vpls-mib.all@tools.ietf.org>, "<mib-doctors@ietf.org>" <mib-doctors@ietf.org>
Subject: Re: [MIB-DOCTORS] Review of draft-ietf-l2vpn-vpls-mib-10
X-BeenThere: mib-doctors@ietf.org
X-Mailman-Version: 2.1.12
Precedence: list
List-Id: MIB Doctors list <mib-doctors.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/mib-doctors>, <mailto:mib-doctors-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/mib-doctors>
List-Post: <mailto:mib-doctors@ietf.org>
List-Help: <mailto:mib-doctors-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/mib-doctors>, <mailto:mib-doctors-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 15 Aug 2013 20:36:59 -0000

"Other" is a good idea too.

Steve

On 8/15/2013 1:23 AM, Giles Heron (giheron) wrote:
> sounds sensible.
>
> There's also a valid option of having static configuration of the PW labels rather than using LDP or BGP signalling, so the enumerated integer probably needs values for "None", "LDP" and "BGP".
>
> Giles
>
> On 15 Aug 2013, at 06:13, Steve Waldbusser <waldbusser@nextbeacon.com> wrote:
>
>> That's exactly what I was thinking (enumerated integer to be more exact). And it could be theoretically extended in the future if other signalling types are invented.
>>
>> Steve
>>
>> On 8/14/2013 4:35 PM, Agrahara Kiran Koushik wrote:
>>> Thanks for your input Steve.
>>>
>>> Do you think adding a new object in the vplsConfigTable to indicate whether BGP or LDP is used for that service
>>> is sufficient to handle this ambiguity?
>>>
>>> Thanks
>>> Kiran
>>>
>>>
>>>>>> 12. Generally: How does the agent know that you intend to start BGP or LDP on a VPLS entry that you
>>>>>> have created with SNMP?
>>>>>>
>>>>> [kiran] Not sure why it would be required?
>>>> Here's what I was thinking:
>>>> Let's say a row is created via SNMP in the vplsConfigTable. Everything the agent knows about that VPLS service is contained in that row. I'm assuming the agent will need to be told whether to use BGP or LDP for that service.
>>>>
>>>> Perhaps it's a side-effect of whether you created an entry in the VplsLdpConfigTable or created one in the vplsBgpConfigTable? Could work I guess but it would be good to prohibit situations where neither entry is created or both are created (especially the latter). Most important would be to define authoritatively how the agent knows. It seemed like the other BDP and LDP tables could be empty even if the corresponding service was enabled so they wouldn't be good choices.
>>>>
>>>> Thanks,
>>>> Steve
>>>>
>>>>
>>>>>
>>>>> Thanks
>>>>> Kiran
>>>>>
>>>>>
>>>>> On Jul 16, 2013, at 8:02 PM, Steve Waldbusser
>>>>> <waldbusser@nextbeacon.com>
>>>>>   wrote:
>>>>>
>>>>>
>>>>>> Hello,
>>>>>>
>>>>>> Here's my review of the the VPLS MIB.
>>>>>>
>>>>>> 1. There are three tables that have a many-to-one relationship with the vplsConfigTable AND that
>>>>>>    allow SNMP create operations. A potential ambiguity exists in the creation process for these
>>>>>>    tables. For example one could create entries in the vplsPwBindTable that reference VPLS entries
>>>>>>    that don't yet exist: IOW one could create vplsPWBindRowStatus.VPLS-A.PW-B and THEN create vpls-A
>>>>>>    in the vplsConfigTable. It is perfectly alright for the inactive rows in the vplsPWBindTable to
>>>>>>    have unresolved dangling references to VPLS-A as long as VPLS-A is created before the PWBind
>>>>>>    Entry is set to active. On the other hand, it has been my experience that agent writers will not
>>>>>>    allow that mode of operation (i.e. will not allow creation of the dangling references) unless the MIB
>>>>>>    explicitly mandates it and I think such a mandate may be more trouble than it's worth. Therefore,
>>>>>>    to protect an NMS writer from ordering their create operations that way and not interoperating,
>>>>>>    I'd suggest explicitly saying that it is optional for the agent to allow entries to be created
>>>>>>    that point to non-existent VPLSConfig entries. The most straightforward path is for the NMS to create the
>>>>>>    VPLSConfig entry and then the vplsPWBind entries.
>>>>>>
>>>>>> The relevant tables are:
>>>>>>
>>>>>>       VplsPwBindEntry ::=
>>>>>>           SEQUENCE {
>>>>>>               vplsPwBindConfigType              INTEGER,
>>>>>>               vplsPwBindType                  INTEGER,
>>>>>>               vplsPwBindRowStatus             RowStatus,
>>>>>>               vplsPwBindStorageType             StorageType
>>>>>>           }
>>>>>>           INDEX  { vplsConfigIndex, pwIndex }
>>>>>>
>>>>>>       VplsBgpRteTargetEntry ::=
>>>>>>          SEQUENCE {
>>>>>>           vplsBgpRteTargetIndex          Unsigned32,
>>>>>>           vplsBgpRteTargetRTType         VplsBgpRouteTargetType,
>>>>>>           vplsBgpRteTargetRT             VplsBgpRouteTarget,
>>>>>>           vplsBgpRteTargetRowStatus      RowStatus
>>>>>>          }
>>>>>>          INDEX     { vplsConfigIndex, vplsBgpRteTargetIndex }
>>>>>>
>>>>>>      VplsBgpVEEntry ::= SEQUENCE {
>>>>>>           vplsBgpVEId          Unsigned32,
>>>>>>           vplsBgpVEName        SnmpAdminString,
>>>>>>           vplsBgpVEPreference  Unsigned32,
>>>>>>           vplsBgpVERowStatus   RowStatus,
>>>>>>           vplsBgpVEStorageType StorageType
>>>>>>         }
>>>>>>         INDEX  { vplsConfigIndex, vplsBgpVEId }
>>>>>>
>>>>>> 2. In a related matter, it would be good to define what happens to rows in the above three tables
>>>>>>    when the associated vplsConfigEntry is deleted. One option is nothing happens (i.e. allow the
>>>>>>    dangling references). Another option is language in the RowStatus objects like: "If this object
>>>>>>    is set to destroy(6) or deleted by the agent, all associated entries in the VplsPWBindTable,
>>>>>>    vplsBGPRteTargetTable and vplsBgpVETable shall be deleted."
>>>>>>
>>>>>>
>>>>>> 3. vplsConfigRowStatus
>>>>>>    vplsPwBindRowStatus
>>>>>>    vplsBgpADConfigRowStatus
>>>>>>    vplsBgpRteTargetRowStatus
>>>>>>    vplsBgpVERowStatus
>>>>>>
>>>>>>    What columns in those tables must be set to valid values before these RowStatus objects can be set
>>>>>>    to active? (Presumably all of them)? [Ref: RFC4181 4.6.4]
>>>>>>
>>>>>> 4. vplsPwBindEntry
>>>>>>    vplsBgpADConfigEntry
>>>>>>    vplsBgpRteTargetEntry
>>>>>>    vplsBgpVEEntry
>>>>>>
>>>>>>    As per RFC4181 4.6.4, "If the agent itself may also create and/or delete rows, then the conditions
>>>>>>    under which this can occur MUST be clearly documented in the row object DESCRIPTION clause."
>>>>>>
>>>>>>    Something like: "Entries in this table may be created or deleted through SNMP, as side-effects of
>>>>>>    console or other non-SNMP management commands, or upon learning via autodiscovery."
>>>>>>
>>>>>> 5. VplsBgpADConfigEntry
>>>>>>    VplsBgpRteTargetEntry
>>>>>>
>>>>>>    These tables need a StorageType object
>>>>>>
>>>>>> 6. vplsPwBindStorageType
>>>>>>    Presumably if this row was from auto-discovery, the storagetype for this row should be 'volatile' or
>>>>>>    'other'. My argument for 'other' is that the entry is 'algorithmically' stored.
>>>>>>    Further, you might want to document that if autodiscovered entries are deleted they would likely
>>>>>>    re-appear in the next autodiscovery interval. I'm guessing you don't intend to
>>>>>>    store a "negative" entry to keep the entry from reappearing. You'd have trouble deleting the
>>>>>>    negative entry anyway.
>>>>>>
>>>>>>    Similar issue for vplsBgpRteTargetRowStatus.
>>>>>>
>>>>>> 7. Page 6: Section 5: Example of the VPLS MIB modules usage
>>>>>>
>>>>>>    Bravo for including an example section!
>>>>>>
>>>>>> 8. This document is structured as three separate MIB modules. Did you consider putting everything in
>>>>>>    one MIB module and using the conformance section to modularize the MIB into 3 optional groups?
>>>>>>
>>>>>> 9. Page 15: vplsStatusTable
>>>>>>
>>>>>>   Do you intend to add a bunch of other stats to this table in the future? Since vplsStatusTable is
>>>>>>   a one-to-one mapping with vplsConfigTable and fate-shares with it, these could just be read-only
>>>>>>   variables in the vplsConfigTable and then this table would be unnecessary. . Further, if it is
>>>>>>   important that this be in a separate table then this should use the AUGMENTS clause, not
>>>>>>   INDEX. AUGMENTS is used for sub-tables that have a one-to-one mapping and that fate-share, like
>>>>>>   this one.
>>>>>>
>>>>>> 10. Page 23 vplsStatusChanged NOTIFICATION-TYPE
>>>>>>
>>>>>>   Need to specify which instances are included in the notification. Something like this: "The object
>>>>>>   instances included in the notification are the ones associated with the VPLS service
>>>>>>   whose status has changed."
>>>>>>
>>>>>>   vplsFwdFullAlarmRaised NOTIFICATION-TYPE
>>>>>>
>>>>>>   Similarly: "The object instances included in the notification are the ones associated with the
>>>>>>   VPLS service which has exceeded the threshold"
>>>>>>
>>>>>>   vplsFwdFullAlarmCleared NOTIFICATION-TYPE
>>>>>>
>>>>>>   Similarly: "The object instances included in the notification are the ones associated with the
>>>>>>   VPLS service which has fallen below the threshold."
>>>>>>
>>>>>> 11. Page 30: vplsLdpConfigMacAddrWithdraw: This object should be read-write, not read-create.
>>>>>>
>>>>>> 12. Generally: How does the agent know that you intend to start BGP or LDP on a VPLS entry that you
>>>>>> have created with SNMP?
>>>>>>
>>>>>> 13. Page 31: vplsLdpPwBindMacAddressLimit
>>>>>>
>>>>>>     Given there's no RowStatus I assume entries in the table are created by the agent as a side
>>>>>>     effect of the creation of PW bindings on an LDP-managed vpls service (regardless of whether the
>>>>>>     service or PW bindings were created manually, by signaling or
>>>>>>     by SNMP.
>>>>>>
>>>>>>     If this assumption is correct then the object should be read-write,
>>>>>>     not read-create.
>>>>>>
>>>>>> 14. Page 42: Security Considerations
>>>>>>
>>>>>>     In order to comply with rfc4181 Section 3.4, this should state explicitly that
>>>>>>     while the read-write and read-create objects must be protected by secure SNMP,
>>>>>>     none of them are especially disruptive. Similarly, while the read-only objects might present
>>>>>>     privacy concerns and due consideration should be given to protecting them with secure SNMP,
>>>>>>     none of these objects contain especially sensitive information.
>>>>>>
>>>>>> I also have a number of smaller items that I wrote notes about when I was on vacation. I need to collect those notes and
>>>>>> send them in another email.
>>>>>>
>>>>>>
>>>>>>
>>>>>> Steve
>>>>>>
>>>>>>

