Re: [i2rs] [yang-doctors] draft-ietf-i2rs-yang-l2-network-topology

[Qin Wu <[e-mail]>]

Good, will see how to tweak on this.

-Qin
发件人: Reshad Rahman (rrahman) [mailto:[e-mail]]
发送时间: 2020年7月9日 10:52
收件人: Qin Wu <[e-mail]>; Susan Hares <[e-mail]>; [e-mail]; [e-mail]
抄送: [e-mail]; 'Benjamin Kaduk' <[e-mail]>
主题: Re: [yang-doctors] draft-ietf-i2rs-yang-l2-network-topology

Hi Qin,

I don’t think this needs feature since it’s pretty common. It’s just IP address (L3) in l2-attributes grouping that I found odd, maybe that grouping should have different name (i.e. not L2 specific). But since I haven’t been involved with this document at all, I probably shouldn’t be making suggestions.

Regards,
Reshad.

From: Qin Wu <[e-mail]<mailto:[e-mail]>>
Date: Wednesday, July 8, 2020 at 10:39 PM
To: "Reshad Rahman (rrahman)" <[e-mail]<mailto:[e-mail]>>, Susan Hares <[e-mail]<mailto:[e-mail]>>, "[e-mail]<mailto:[e-mail]>" <[e-mail]<mailto:[e-mail]>>, "[e-mail]<mailto:[e-mail]>" <[e-mail]<mailto:[e-mail]>>
Cc: "[e-mail]<mailto:[e-mail]>" <[e-mail]<mailto:[e-mail]>>, 'Benjamin Kaduk' <[e-mail]<mailto:[e-mail]>>
Subject: RE: [yang-doctors] draft-ietf-i2rs-yang-l2-network-topology

Reshad:
The management IP address is used by the administrator to get access to L2 device. It is just listed as a node attribute. We could add feature to make it optional if you think needed.

-Qin
发件人: i2rs [mailto:[e-mail]] 代表 Reshad Rahman (rrahman)
发送时间: 2020年7月9日 0:20
收件人: Susan Hares <[e-mail]<mailto:[e-mail]>>; [e-mail]<mailto:[e-mail]>; [e-mail]<mailto:[e-mail]>
抄送: [e-mail]<mailto:[e-mail]>; 'Benjamin Kaduk' <[e-mail]<mailto:[e-mail]>>
主题: Re: [i2rs] [yang-doctors] draft-ietf-i2rs-yang-l2-network-topology

Thanks Sue. I wasn’t questioning the need for multiple IP addresses, I just found it odd to have IP address(es) in an L2 grouping.

Regards,
Reshad.

From: Susan Hares <[e-mail]<mailto:[e-mail]>>
Date: Wednesday, July 8, 2020 at 12:16 PM
To: "Reshad Rahman (rrahman)" <[e-mail]<mailto:[e-mail]>>, "[e-mail]<mailto:[e-mail]>" <[e-mail]<mailto:[e-mail]>>, "[e-mail]<mailto:[e-mail]>" <[e-mail]<mailto:[e-mail]>>
Cc: "[e-mail]<mailto:[e-mail]>" <[e-mail]<mailto:[e-mail]>>, 'Benjamin Kaduk' <[e-mail]<mailto:[e-mail]>>
Subject: RE: [yang-doctors] draft-ietf-i2rs-yang-l2-network-topology

Reshad:

Juregen indicates my questions are not for IETF but for IEEE, and that multiple system-macs may be implemented as a list.

I’m looking at the IEEE  yang model for 802-1Qcp (2018 update to IEEE 802.1Q), but you can get the pre-release draft at (only editorial nits are missing) at:

http://www.ieee802.org/1/files/private/cp-drafts/d2/802-1Qcp-d2-2.pdf

The reason there are multiple IP addresses is that the single LAN can have multiple IP addresses assigned to it.
In may routers/switches the single MAC port will have multiple virtual ports.  (cisco started this concept in the 1990s).

Thanks for the help!

Sue

From: Reshad Rahman (rrahman) [mailto:[e-mail]]
Sent: Wednesday, July 8, 2020 12:09 PM
To: Susan Hares; [e-mail]<mailto:[e-mail]>; [e-mail]<mailto:[e-mail]>
Cc: [e-mail]<mailto:[e-mail]>; 'Benjamin Kaduk'
Subject: Re: [yang-doctors] draft-ietf-i2rs-yang-l2-network-topology

Hi,

I was referring to management-address leaf-list in the same grouping (see below).

I don’t have the expertise to answer your questions below, maybe some YDs do.

Regards,
Reshad.

     grouping l2-node-attributes {
       description
         "L2 node attributes";
       container l2-node-attributes {
         description
           "Contains L2 node attributes.";
<snip>
         leaf-list management-address {
           type inet:ip-address;
           description
             "System management address.";
         }
         leaf sys-mac-address {
           type yang:mac-address;
           description
             "System MAC address.";
         }


From: Susan Hares <[e-mail]<mailto:[e-mail]>>
Date: Wednesday, July 8, 2020 at 11:48 AM
To: "Reshad Rahman (rrahman)" <[e-mail]<mailto:[e-mail]>>, "[e-mail]<mailto:[e-mail]>" <[e-mail]<mailto:[e-mail]>>, "[e-mail]<mailto:[e-mail]>" <[e-mail]<mailto:[e-mail]>>
Cc: "[e-mail]<mailto:[e-mail]>" <[e-mail]<mailto:[e-mail]>>, 'Benjamin Kaduk' <[e-mail]<mailto:[e-mail]>>
Subject: RE: [yang-doctors] draft-ietf-i2rs-yang-l2-network-topology

Reshad:

Thank you for quick response.   However, I’m confused.   Where do you see the IP address in the Yang snippet I sent?  It is a MAC Address (type yang:mac-address)?  If the system port allows IP, it will respond to the ARP request with the appropriate IP/MAC match in an ARP reply.

In real implementations I2rs implementers examined,

  *   L2 port with mgmt-mac-address seems to be used to only send LLDP packets,
  *   L2 port with sys-mac-address seems to be used for management using TCP/IP.



If we swap the names, it did not work with the initial implementers of the yang model.  The  I2RS Topology models are used for operational management of switches and routers as logical units.



Also, your response does not seem to match my questions:


1) Is this the normal assumption for yang models?
2) If not, what is the normal assumption on system mac addresses?
3) Am I correct that switches with more than 1 system MAC will augment their basic yang model with the second system MAC Address?

I’m sorry to bother you but this document is being reviewed by the IESG  tomorrow (Thursday) and
I am the shepherd.  I do not know how to answer some of the yang related questions regarding multiple system ports to exchange management configuration on.

I had thought that netconf/restconf would be exchanged over the same ports so the yang doctors would know what the normal custom should be.

Thanks again!

Sue










From: Reshad Rahman (rrahman) [mailto:[e-mail]]
Sent: Wednesday, July 8, 2020 11:27 AM
To: Susan Hares; [e-mail]<mailto:[e-mail]>; [e-mail]<mailto:[e-mail]>
Cc: [e-mail]<mailto:[e-mail]>; 'Benjamin Kaduk'
Subject: Re: [yang-doctors] draft-ietf-i2rs-yang-l2-network-topology

Hi,

So sys-mac-address is supposed to be the MAC address of the mgmt port, i.e. the device would respond to an ARP request for management-address with sys-mac-address? I think use of term system might be a bit misleading if that’s the case, mgmt-mac-address might be better.

Also, it is odd to have an IP address in an L2 grouping.

Disclaimer: not familiar with that draft at all, just took a look at the L2 grouping.

Regards,
Reshad.

From: yang-doctors <[e-mail]<mailto:[e-mail]>> on behalf of Susan Hares <[e-mail]<mailto:[e-mail]>>
Date: Wednesday, July 8, 2020 at 11:12 AM
To: "[e-mail]<mailto:[e-mail]>" <[e-mail]<mailto:[e-mail]>>, "[e-mail]<mailto:[e-mail]>" <[e-mail]<mailto:[e-mail]>>
Cc: "[e-mail]<mailto:[e-mail]>" <[e-mail]<mailto:[e-mail]>>, 'Benjamin Kaduk' <[e-mail]<mailto:[e-mail]>>
Subject: [yang-doctors] draft-ietf-i2rs-yang-l2-network-topology

The following question was asked by Ben Kaduk during IESG review of the following document:

https://datatracker.ietf.org/doc/draft-ietf-i2rs-yang-l2-network-topology/


     grouping l2-node-attributes {

         [...]

         leaf sys-mac-address {

           type yang:mac-address;

           description

             "System MAC address.";

         }

If there are more than 1 system mac address in a switch, how would this model handle it.

My understanding is that most switches have 1 system mac address for network management.  Therefore, the L2 topology model supports one.

Question for Yang Doctors:

1) Is this the normal assumption for yang models?
2) If not, what is the normal assumption on system mac addresses?

3) Am I correct that switches with more than 1 system MAC will augment their basic yang model with the second system MAC Address.

Thank you, Susan Hares