Re: [Netconf] Draft charter update for review

[Rodney Cummings <[e-mail]>]

Hi Benoit,

I think we will need to agree to disagree on this issue.

To me, the sentence in the charter sounds like the sort of thinking that 
resulted in XML being mandated for RESTCONF. After all, if NETCONF 
mandates XML, one can argue that RESTCONF should not deviate. 

Did the working group provide a "proper justification" to remove the XML 
mandate from RESTCONF?

I don't know, because the phrase "proper justification" is subjective. 
Maybe we did, and maybe we didn't. Therefore, it seems likely that we will 
repeat this sort of debate. Someone will propose to mandate a feature from 
NETCONF that increases the RESTCONF implementation burden. We can have 
another lengthy debate and vote each time this happens, but I don't see 
that as a good use of everyone's time.

If the sentence was changed to 
        "RESTCONF should provide optional features that align with NETCONF 
capabilities, unless proper justification is provided and documented."
then I would have no concerns. Addition of optional features does not 
increase the implementation complexity, because a server an choose to 
ignore the optional features.

Rodney



From:   Benoit Claise <[e-mail]>
To:     Rodney Cummings <[e-mail]>, "Ersue, Mehmet (Nokia - 
DE/Munich)" <[e-mail]>, 
Cc:     Netconf <[e-mail]>
Date:   09/30/2015 06:05 AM
Subject:        Re: [Netconf] Draft charter update for review



Hi Rodney,

The sentence you mention tells: NETCONF and RESTCONF may differ but only 
for a good and documented reason.
And I believe it does the job.

Regards, Benoit
Hello, 

I apologize for commenting a bit late on the proposed charter, but I have 
a concern. 

The proposal states: 
> RESTCONF should not deviate from the NETCONF capabilities unless proper 
justification 
> is provided and documented. 

RESTCONF is currently being explored for use in many applications due to 
the fact that it is a subset of NETCONF capabilities, with a simpler 
implementation in servers. This is described in section 1.1 of the current 
draft (i.e. "Simple Subset of NETCONF Functionality"). 

In my opinion this is a benefit of RESTCONF that cannot be lost. If 
anything, the opposite strategy should be taken, in that NETCONF 
capabilities should only be added to RESTCONF with a strong justification. 


Rodney 




From:        "Ersue, Mehmet (Nokia - DE/Munich)" <[e-mail]> 
To:        Netconf <[e-mail]>, 
Date:        09/17/2015 05:25 PM 
Subject:        [Netconf] Draft charter update for review 
Sent by:        "Netconf" <[e-mail]> 



Dear NETCONF WG, 
  
please find below the draft charter update which we provided to our AD for 
review. 
Comments are welcome. 
  
Mehmet & Mahesh 
  
------------------   
  
Network Configuration (netconf) 
------------------------------- 
  
Charter 
  
Current Status: Active 
  
Chairs: 
    Mahesh Jethanandani <[e-mail]> 
    Mehmet Ersue <[e-mail]> 
  
Operations and Management Area Directors: 
     Benoit Claise <[e-mail]> 
     Joel Jaeggli <[e-mail]> 
  
Operations and Management Area Advisor: 
     Benoit Claise <[e-mail]> 
  
Mailing Lists: 
     General Discussion: [e-mail] 
     To Subscribe:       https://www.ietf.org/mailman/listinfo/netconf 
     Archive:            https://mailarchive.ietf.org/arch/browse/netconf/ 

  
Description of Working Group: 
  Configuration of networks of devices has become a critical requirement 
  for operators in today's highly interconnected networks. Large and small 

  operators alike have developed their own mechanisms or have used vendor 
  specific mechanisms to transfer configuration data to and from a device 
  and to examine device state information which may impact the 
configuration. Each of these mechanisms may be different in various 
  aspects, such as session establishment, user authentication, 
  configuration data exchange, and error responses. 
  
  The NETCONF protocol (RFC 6241) provides mechanisms to install, 
manipulate, 
  and delete the configuration of network devices. NETCONF is based on the 

  secure transport (SSH is mandatory to implement while TLS is an optional 

  transport) and uses an XML-based data representation. The NETCONF 
protocol 
  is data modeling language independent, but YANG (RFC 6020) is the 
recommended 
  NETCONF modeling language, which introduces advanced 
  language features for configuration management. 
  
  Recently NETCONF WG define the Call Home mechanism for NETCONF and 
RESTCONF protocols 
  as well as updated RFC5539 by adding the new message framing used by 
NETCONF 1.1. 
  
  Based on the implementation, deployment experience and interoperability 
  testing, the WG aims to produce a NETCONF status report in a later 
stage. 
  The result may be clarifications for RFC6241 and RFC6242 and addressing 
  any reported errata. 
  
  In the current phase of NETCONF's incremental development the workgroup 
  will focus on following items: 
  
  1. Combine the server configuration data models from the Call Home and 
RFC5539bis drafts in a separate Server Configuration YANG module by 
addressing the need for NETCONF as well as RESTCONF. 
  
  2. Develop RESTCONF, a protocol based on NETCONF in terms of 
capabilities, but over HTTP and with some REST characteristics, for 
accessing YANG data in NETCONF datastores. An "ordered edit list" approach 
is needed (the YANG patch) to provide client developers with a simpler 
edit request format that can be more efficient and also allow more precise 
client control of the transaction procedure than existing mechanisms. The 
YANG patch operation, based on the  HTTP PATCH method, will be prepared in 
a separate draft. 
In addition develop a YANG library, which identifies the information about 
all YANG modules used by a server. Furthermore develop a collection 
resource for the RESTCONF protocol to provide enhanced filtering features 
for the retrieval of data nodes with the GET method. 
RESTCONF should not deviate from the NETCONF capabilities unless proper 
justification is provided and documented. The RESTCONF work will consider 
requirements suggested by the other working groups (for example I2RS). 
  
    3. Develop a zero touch configuration document (a technique to 
establish a secure network management relationship between a newly 
delivered network device configured with just its factory default 
settings, and the Network Management System), specific to the NETCONF use 
case. 
  
    4. Develop a subscription and push mechanism that allows client 
applications to request notifications for changes in the datastore. These 
updates will be pushed by the server to the client based on a subscription 
policy. 
  
    5. Update RFC 6536 (NETCONF Access Control Model) to introduce access 
control rights associated with actions. 
  
    6.Enhance RFC 5277 with the ability to delete subscriptions without 
closing the client session, to modify existing subscriptions, and to have 
multiple subscriptions on a established client session. These changes 
should not affect older clients that do not support these particular 
subscription requirements. The RPCs and the data models in RFC 5277 should 
be converted to YANG. 
  
Goals and Milestones: 
  
  Done - Submit RFC5539bis to AD/IESG for consideration as Proposed 
Standard 
  Done - Submit NETCONF call home mechanism to AD/IESG for consideration 
as Proposed Standard 
  
  Oct 2015 - WGLC for RESTCONF, YANG patch operation and YANG Library 
drafts 
  Nov 2015 - Submit RESTCONF to AD/IESG for consideration as Proposed 
Standard 
  
  Jan 2016 - WGLC for RFC5277bis draft 
  Jan 2016 - Submit RFC5277bis to AD/IESG for consideration as Proposed 
Standard 
  Jan 2016 - WGLC for YANG datastore push update draft 
  Feb 2016 - Submit YANG datastore push update to AD/IESG for 
consideration as Proposed Standard 
  
  Feb 2016 - WGLC for RFC6536bis draft 
  Feb 2016 - Submit RFC6536bis to AD/IESG for consideration as Proposed 
Standard 
  
  Feb 2016 - WGLC for zero touch configuration 
  Feb 2016 - WGLC for RESTCONF Collection Resource draft 
  Mar 2016 - Submit zero touch configuration to AD/IESG for consideration 
as Proposed Standard 
  Mar 2016 - Submit RESTCONF Collection to AD/IESG for consideration as 
Proposed Standard 
  
  Mar 2016 - WGLC for NETCONF server configuration data model 
  Apr 2016 - Submit server configuration data model to AD/IESG for 
consideration as 
  
 _______________________________________________
Netconf mailing list
[e-mail]
https://www.ietf.org/mailman/listinfo/netconf


_______________________________________________
Netconf mailing list
[e-mail]
https://www.ietf.org/mailman/listinfo/netconf