Re: [vnfpool] FW: New Version Notification for draft-zong-vnfpool-problem-statement-02.txt

Hi Ning,
1- I understand your point, i.e., "/... not to address the problem from 
service(or user) perspective, but simply from the VNF set itself/.". 
Actually, the VNF set reliability is commonly related to its 
availability, and its availability impacts on the service (which uses 
this VNF set) availability.
Nevertheless, a "purist" could say the following: '/the VNF set is 
reliable, i.e., once it starts working, you can trust it, because its 
embedded mechanisms such as redundancy will always provide a non-stop 
service BUT it may not have a high availability because, sometimes, when 
you need to launch the VNF set, it does not start .../'
Well, beyond this digression, ok for me to just mention that this "/work 
will benefit service availability/".
2- /performance /vs. /reliability /for load sharing (LS): I still think 
that LS is part of resource optimization mechanisms (if N components can 
do the same job, let's share the requests for this type of job properly 
- for efficiency purposes) and then, as a _second_ gain/advantage, LS 
helps to improve the overall reliability. Again, it is a subjective 
opinion ... It is like a half empty/full glass ...
Thank you + best,
Chidung

Le 18/01/2014 02:31, Zongning a écrit :
>
> Hi, Chidung,
>
> Really appreciate your inline suggestions + comments. While we will 
> adopt most of your suggestions, I'd like to answer several of your 
> questions at first, to make sure we are more align with each other. J
>
> [CL1]: Besides section 5 (Related Works), the word « availability »  
> is not used at all in this doc ... although it is the main objective 
> of the pooling architecture, i.e., providing redundant VNF instances 
> in order to maximize the network service's availability
> You are right that the goal of VNF pool is to improve the service's 
> availability, from the service(or user) perspective. IMHO, I prefer 
> not to address the problem from service(or user) perspective, but 
> simply from the VNF set itself. Another word, we may just focus how to 
> use pooling mechanisms, e.g. redundancy model, stateful failover to 
> improve the reliability of VNF set. But we could definitely mention 
> that our work will benefit service availability. How do you think?
>
> [CL2]: "A VNF set can introduce additional points of failure beyond 
> thoseinherent in a single specialized server" - this sentence implies 
> that a "single specialized server" currently hosts different NFs which 
> will be virtualized in the future and  grouped in a "VNF set": is it 
> really what we mean ?
>
> No, we don't mean that VNF set is "released" from a single specialized 
> server. The challenges are: 1) a single VNF may have more factors of 
> risk than a traditional NF in single specialized server; 2) a group of 
> VNFs need more coordination on reliability. I agree that comparing a 
> VNF set to a single specialized server may be inappropriate. I will 
> revise the sentence later.
>
> [CL4]: "Load Sharing" - it is rather a performance mechanism (and not 
> a reliability one).
>
> I'd still think load sharing is not only for performance, but related 
> to reliability, considering that potential overload will affect the 
> reliability of VNF without applying efficient load sharing.
>
>
> Again, thanks for your careful & kind inline editing.
>
> -Ning
>
> *From:*LAC Chidung [mailto:chidung.lac@orange.com]
> *Sent:* Saturday, January 18, 2014 12:53 AM
> *To:* Zongning
> *Cc:* vnfpool@ietf.org
> *Subject:* Re: [vnfpool] FW: New Version Notification for 
> draft-zong-vnfpool-problem-statement-02.txt
>
> Hi Ning,
> Some cosmetic suggestions + comments.
> Best,
> Chidung
>
> Le 14/01/2014 02:24, Zongning a écrit :
>
>     Hi, folks,
>
>     Happy new year!
>
>     We made a new revision of VNF Pool Problem Statement. Please see the below abstract and links:
>
>       
>
>     Abstract:
>
>         There is a trend to implement network services by using a group of
>
>         Virtualized Network Functions (VNFs) called a VNF set.  A VNF set can
>
>         offer benefits like flexible service provisioning.  A VNF set also
>
>         introduces additional points of failure, and therefore poses
>
>         additional challenges on reliability.
>
>       
>
>         This document overviews the problems related to the reliability of a
>
>         VNF set.  A VNF pooling architecture is also briefly introduced as
>
>         candidate solution to address the problems.
>
>       
>
>     URL:http://www.ietf.org/internet-drafts/draft-zong-vnfpool-problem-statement-02.txt
>
>     Htmlized:http://tools.ietf.org/html/draft-zong-vnfpool-problem-statement-02
>
>       
>
>     Other related drafts including VNF Pool Use Cases, VNF Pooling Architecture are under development and will be available soon in this month...
>
>     Thanks.
>
>     -Ning
>
>       
>
>         -----Original Message-----
>
>         From:internet-drafts@ietf.org  <mailto:internet-drafts@ietf.org>  [mailto:internet-drafts@ietf.org]
>
>         Sent: Tuesday, January 14, 2014 9:18 AM
>
>         To: Zongning; Linda Dunbar; Melinda Shore; Linda Dunbar; Melinda Shore;
>
>         Diego Lopez; Diego R. Lopez; Zongning
>
>         Subject: New Version Notification for
>
>         draft-zong-vnfpool-problem-statement-02.txt
>
>           
>
>         A new version of I-D, draft-zong-vnfpool-problem-statement-02.txt
>
>         has been successfully submitted by Ning Zong and posted to the
>
>         IETF repository.
>
>           
>
>         Name:             draft-zong-vnfpool-problem-statement
>
>         Revision: 02
>
>         Title:            Virtualized Network Function (VNF) Pool Problem Statement
>
>         Document date:    2014-01-14
>
>         Group:            Individual Submission
>
>         Pages:            10
>
>         URL:
>
>         http://www.ietf.org/internet-drafts/draft-zong-vnfpool-problem-statement-02.
>
>         txt
>
>         Status:
>
>         https://datatracker.ietf.org/doc/draft-zong-vnfpool-problem-statement/
>
>         Htmlized:
>
>         http://tools.ietf.org/html/draft-zong-vnfpool-problem-statement-02
>
>         Diff:
>
>         http://www.ietf.org/rfcdiff?url2=draft-zong-vnfpool-problem-statement-02
>
>           
>
>         Abstract:
>
>             There is a trend to implement network services by using a group of
>
>             Virtualized Network Functions (VNFs) called a VNF set.  A VNF set can
>
>             offer benefits like flexible service provisioning.  A VNF set also
>
>             introduces additional points of failure, and therefore poses
>
>             additional challenges on reliability.
>
>           
>
>             This document overviews the problems related to the reliability of a
>
>             VNF set.  A VNF pooling architecture is also briefly introduced as
>
>             candidate solution to address the problems.
>
>           
>
>         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.
>
>         The IETF Secretariat
>
>     _______________________________________________
>
>     vnfpool mailing list
>
>     vnfpool@ietf.org  <mailto:vnfpool@ietf.org>
>
>     https://www.ietf.org/mailman/listinfo/vnfpool
>
> ------------------------------------------------------------------------
>