[Ibnemo] 答复: [Sdn] I-D Action: draft-xia-sdnrg-service-description-language-02.txt

[Xiayinben <[e-mail]>]

Hi Albrecht,

You are right, this draft is focus on the requirement that simplify to describe user's intent of manipulating network. 
I think a language, especially DSL(domain specific language), is user friendly way to describe user's intent. So we suggest to use a description language in this draft.
The example in this draft is using to illustrate user's intent rather than language design.

We have another draft is focus on the language design, https://tools.ietf.org/html/draft-xia-sdnrg-nemo-language-02. I hope you are interesting too. And we are looking forward to your comments. 
As you indicated, policy is in form of "if conditions then actions". It always be used in response to some change situation. But most time user has direct command to do something, e.g. provisioning a virtual network. So I think policy is a part of description language for user's intent, but it is not all.

Best regards,
Yinben 

-----邮件原件-----
发件人: sdn [mailto:[e-mail]] 代表 Schwarz, Albrecht (Albrecht)
发送时间: 2015年5月13日 21:02
收件人: Zhoutianran; STUART VENTERS; [e-mail]
抄送: [e-mail]
主题: Re: [Sdn] I-D Action: draft-xia-sdnrg-service-description-language-02.txt

Hello Terence,

the scope of your draft covers two topics:
a) requirements for a language and
b) design aspects ("considerations") of a language.

The purpose of this draft is NOT to specify a new language in my understanding, right?

Languages for the description of (commmunication services) are of type "system design language" (using the notion of ITU-T).
(Z.119, 3.3  system design language: A formal notation that has a well-defined syntax and semantics that is used for some aspect of the engineering systems ranging from statements of requirements for procurement or in standards to the deployment of implemented systems or networks.)

These are formal languages, - whereas your draft so far seems to consider informal language style, if not, then there should be a requirement that the language should support a formal notation...

Regards,
Albrecht

PS
I would state that UML could be the solution in the end to your draft, - or are there any "service description" requirements which can't be satisfied by UML? Don't think so.

PSS
Your examples in clause 4 indicating a specific service aspect only, which fairly overlaps with policy rule control, that's why the PSL side discussion. 


-----Original Message-----
From: Zhoutianran [mailto:[e-mail]]
Sent: Dienstag, 12. Mai 2015 08:06
To: Schwarz, Albrecht (Albrecht); STUART VENTERS; [e-mail]
Cc: [e-mail]
Subject: RE: [Sdn] I-D Action: draft-xia-sdnrg-service-description-language-02.txt

Hi Albrecht,

I agree with you on the second. There are more options. And you gave many examples.
I do not know what you mean for the first comment. Could you give more information?


Best Regards,
Terence


-----Original Message-----
From: Schwarz, Albrecht (Albrecht) [mailto:[e-mail]]
Sent: Saturday, May 09, 2015 12:00 AM
To: Zhoutianran; STUART VENTERS; [e-mail]
Cc: [e-mail]
Subject: RE: [Sdn] I-D Action: draft-xia-sdnrg-service-description-language-02.txt

Terence, et al.,

firstly, the level of languages and level of protocols should be not mixed in the ietf draft.

Secondly, there are even more options:

I) Compilation/interpretation dimension:
O1: in runtime evironment
O2: in advance

II) Signalling (= protocol) dimension (with x= 1, 2):
Ox.1: policy rule explicitly signalled
Ox.2: pointer to policy rule signaled, policy rule then retrieved from a remote, central data base by enforcement element
Ox.3: pointer to policy rule signaled, policy rule retrieved from a local data base

Furthermore:
Ox.1.1: policy rule signaled in a container, i.e., PSL syntax is different to underlying protocol syntax
Ox.1.2: policy rule signaled in native protocol syntax

Regards,
Albrecht

-----Original Message-----
From: sdn [mailto:[e-mail]] On Behalf Of Zhoutianran
Sent: Freitag, 8. Mai 2015 04:56
To: STUART VENTERS; [e-mail]
Cc: [e-mail]
Subject: Re: [Sdn] I-D Action: draft-xia-sdnrg-service-description-language-02.txt

This is interesting. Here is my two cents.

I agree with you that there are two issues.
And for the first one, I think we need a language to easily express the service request or the user intent.

The second one is about the protocol between the controller and applications. I think we have two option.
1. One is to pass the script as content directly to the controller. And all the compilation/parsing done on the controller.
2. The other option is to parse the intent expression on the application side into some well known format, e.g. xml, json, and then transport the content to the controller.

That's what we need to discuss in the community.


Best,
Terence


-----Original Message-----
From: sdn [mailto:[e-mail]] On Behalf Of STUART VENTERS
Sent: Wednesday, May 06, 2015 11:19 PM
To: [e-mail]
Subject: Re: [Sdn] I-D Action: draft-xia-sdnrg-service-description-language-02.txt

There appear to be two, separable technical issues in this draft:

1) What information is needed to represent a service request.
2) How to represent and transport the information for a request.

The first part may be a good area.

For the second part, I wonder if there are not already available methods to declare a schema and represent and transport the information governed by it.
It appears the draft is proposing to invent another mechanism.
It would be nice to understand the reasoning behind this choice.


--------------

On 04-05-15 11:59, [e-mail] wrote:
>
> A New Internet-Draft is available from the on-line Internet-Drafts directories.
>
>
>          Title           : Requirements for a Service Description Language and Design Considerations
>          Authors         : Yinben Xia
>                            Sheng Jiang
>                            Susan Hares
> 	Filename        : draft-xia-sdnrg-service-description-language-02.txt
> 	Pages           : 7
> 	Date            : 2015-05-04
>
> Abstract:
>     The more and more complicated IP networks require a new interaction
>     mechanism between their customers and their networks.  A service
>     description language is needed to enable customers to easily describe
>     their diverse intent.  SDN controller would compile the user intent
>     into device configurations.  This document analyzes requirements for
>     such service description language and gives considerations for
>     designing such service description language.
>
>
> The IETF datatracker status page for this draft is:
> https://datatracker.ietf.org/doc/draft-xia-sdnrg-service-description-l
> anguage/
>
> There's also a htmlized version available at:
> https://tools.ietf.org/html/draft-xia-sdnrg-service-description-langua
> ge-02
>
> A diff from the previous version is available at:
> https://www.ietf.org/rfcdiff?url2=draft-xia-sdnrg-service-description-
> language-02
>
>
> Please note that it may take a couple of minutes from the time of 
> submission until the htmlized version and diff are available at tools.ietf.org.
>
> Internet-Drafts are also available by anonymous FTP at:
> ftp://ftp.ietf.org/internet-drafts/
>
> _______________________________________________
> I-D-Announce mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/i-d-announce
> Internet-Draft directories: http://www.ietf.org/shadow.html or 
> ftp://ftp.ietf.org/ietf/1shadow-sites.txt
>


--
*****************************************************************
               请记住，你是独一无二的，就像其他每一个人一样

_______________________________________________
sdn mailing list
[e-mail]
https://www.irtf.org/mailman/listinfo/sdn
_______________________________________________
sdn mailing list
[e-mail]
https://www.irtf.org/mailman/listinfo/sdn
_______________________________________________
sdn mailing list
[e-mail]
https://www.irtf.org/mailman/listinfo/sdn
_______________________________________________
sdn mailing list
[e-mail]
https://www.irtf.org/mailman/listinfo/sdn