[vnfpool] 答复: comments on VNF Pool Charter

[Zongning <[e-mail]>]

Hi, Kapil,

Please see inline.

发件人: Sood, Kapil [mailto:[e-mail]]
发送时间: 2014年6月18日 6:44
收件人: Zongning; [e-mail]
主题: RE: comments on VNF Pool Charter

Hello Ning,

Thank you for your comments.  Please see below:

Best Regards,

Kapil.

From: Zongning [mailto:[e-mail]]
Sent: Monday, June 16, 2014 7:16 PM
To: Sood, Kapil; [e-mail]<mailto:[e-mail]>
Subject: 答复: comments on VNF Pool Charter

Hi, Kapil,

Thanks for the review. Please see inline.

Hello All,

A couple comments, maybe questions, on the VNF Charter:

1.      The point below indicates “to collect network information”…did you consider collecting the platform information as well?  I would suggest to add that.

[Ning] I’d think “platform information” may be too general term to be focused. In our charter, we collect network information for specific purpose, e.g., place a live and backup VNF instances into distributed physical locations, or locations without shared risk in the network. Could you please elaborate more on the purpose of using any specific piece of platform information, or give some use cases? We could then think about how to make it more focused, as required by the charter.
[Kapil] ETSI-NFV has defined multiple use cases for NFV placements/migration that depends on the target NFVI (platform) conditions – e.g. available CPU, Memory, Network Capabilities, etc.  So, it makes sense to consider those requirements as we define this charter.  Please share what network information this group is looking to collect and what are the possible sources?  Those might not be too different from some of the VNFI information sources.

[Ning] You are right that a rich set of platform conditions are useful for VNF instance placement. This group will not address the whole issue of VNF deployment, but only focus on redundancy. One issue we have identified in Problem Statement I-D is to distribute active & standby instances into different network locations to avoid shared risk. Some existing interfaces defined by IETF (e.g., ALTO, I2RS) may be utilized to obtain network information such as network topology, link status, for such purpose. I do agree with you that if there are more requirements on placing VNF instance, more NFVI information should be considered. At this stage, I’d prefer to add a few words like “considering the infrastructure conditions…” in the bullet under “Questions that are raised…” in the charter. Is this OK for you?


2.      Please clarify the scope of the protocols work within the charter.  Current text says “re-chartering before adopting any work to develop new, or extend existing, protocols”.  But, it also says that group will work on 2 protocols.

[Ning] Current charter lists a set of interfaces/protocols to be standardized, but will only delivers documents such as requirements to these interfaces/protocols, and architecture highlighting the functionalities of these interfaces/protocols. The protocol development, i.e., specifying on-the-wire message format, will be after the re-chartering.
[Kapil] OK, thanks.

Best Regards,

Kapil.

Hope my reply helps.

Thanks.

-Ning