Re: [vnfpool] Overlap with SFC?

[LAC Chidung <[e-mail]>]

_@ Linda:_ "/If a service function has a small number of instances, say 
less than 5/": a service function (SF) is composed of 5 Network 
Functions (NFs) - correct ?
_@ Susan:_ if the answer to the question above is "Y", then "/You may 
have parallel chains to support the redundancy/" seems strange ... 
Actually, a chain characterizes a SF, i.e., to obtain a certain SF, we 
are putting some NFs (5NFs in the previous example) in a chain. If this 
SF has to be highly available, the need is to make each NF highly 
available (using vnfpool for instance), but not to create another chain, 
because there is one unique chain to render the SF we are currently 
designing.
Thank you for correcting my interpretation.
Best,
Chidung

Le 25/01/2014 02:04, Susan Hares a écrit :
>
> Linda:
>
> I agree with your use example, but would like to clarify it.
>
> VNF Pools  is about redundancy needs rather than chaining paths.  
>  Pools provide the redundancy.   You may need the redundancy for even 
> a short chain if it is High-availability.   You may have parallel 
> chains to support the redundancy.
>
> Cheers,
>
> Sue
>
> *From:*vnfpool [mailto:[e-mail]] *On Behalf Of *Linda 
> Dunbar
> *Sent:* Friday, January 24, 2014 5:55 PM
> *To:* Qin Wu; LAC Chidung
> *Cc:* [e-mail]; [e-mail]
> *Subject:* Re: [vnfpool] Overlap with SFC?
>
> Here is my take of the differences and the overlap with SFC:
>
> If a service function has a small number of instances, say less than 
> 5, and they are relative stable, then it is doable for Service Chain 
> to specify the specific instances, the so called "Service Chain Path" 
> by SFC drafts.
>
> But if there are large number of instances, say in hundreds, and those 
> instances' location/presence change over time (e.g. in NFV 
> environment), then it is not scalable to have Service Chain path to 
> specify the specific instances.
>
> The service function instances' management, e.g. selection, 
> replacement when failure occurs or over-utilized from a pool of 
> available instances, notification should be  the scope of VNFpool.
>
> Linda
>
> *From:*vnfpool [mailto:[e-mail]] *On Behalf Of *Qin Wu
> *Sent:* Wednesday, January 22, 2014 7:38 PM
> *To:* LAC Chidung
> *Cc:* [e-mail] <mailto:[e-mail]>; [e-mail] 
> <mailto:[e-mail]>
> *Subject:* Re: [vnfpool] Overlap with SFC?
>
> Hi, Chidung:
>
> That's one good interpretation.
>
> You are right, I should add assumption that the service node that is 
> bypassed is optional service node in the chain.
>
> Regards!
>
> -Qin
>
> *From:*vnfpool [mailto:[e-mail]] *On Behalf Of *LAC 
> Chidung
> *Sent:* Wednesday, January 22, 2014 7:26 PM
> *To:* Qin Wu
> *Cc:* [e-mail] <mailto:[e-mail]>; [e-mail] 
> <mailto:[e-mail]>
> *Subject:* Re: [vnfpool] Overlap with SFC?
>
> Hi Qin,
> "/in service chain, when a service node is down, what SFC is doing is 
> to bypass this service node/": is the following interpretation ok ?
> A service chain needs X+Y service nodes in order to provide the 
> *_full_* service: the X service nodes are mandatory, while the Y 
> service nodes are (kind of) optional, i.e., if one (or more) of these 
> Y service nodes is (are) down, the service chain can still provide the 
> service, but in a *_degraded mode_*. In this case,*the bypassing can 
> only happen* for one of the Y service nodes, i.e., if one of the X 
> service nodes is down, there is *_no service at all_*.
> _NB:_ in this example, we consider, of course, that there is no 
> redundancy anywhere, i.e., if a service node is down, the only thing 
> to do is to fix it, and while waiting for the reparation, we face a 
> degraded service, or no service at all.
> Best,
> Chidung
>
> Le 22/01/2014 06:05, Qin Wu a écrit :
>
>     My Understanding is
>
>     in service chain, when a service node is down, what SFC is doing is to bypass this service node while
>
>     What nfvpool is doing is to replace the failing one with the new service node which provide the same functionality.
>
>     Please correct me if I am wrong.
>
>     Regards!
>
>     -Qin
>
>     -----Original Message-----
>
>     From: vnfpool [mailto:[e-mail]] On Behalf Of Zongning
>
>     Sent: Wednesday, January 22, 2014 9:42 AM
>
>     To: Zongning;[e-mail]  <mailto:[e-mail]>;[e-mail]  <mailto:[e-mail]>
>
>     Subject: Re: [vnfpool] Overlap with SFC?
>
>       
>
>     Sorry, item 1) is obviously not finished. :-)
>
>     1) SFC targets on steering packets among service function nodes. vnfpool focuses on redundancy for service nodes, e.g., selecting standby nodes, handling nodes transition/failure cases, without caring about how to construct the data path.
>
>     Again, my fault.
>
>     -Ning
>
>       
>
>         -----Original Message-----
>
>         From: Zongning
>
>         Sent: Wednesday, January 22, 2014 9:34 AM
>
>         To: '[e-mail]  <mailto:[e-mail]>';[e-mail]  <mailto:[e-mail]>
>
>         Subject: RE: [vnfpool] Overlap with SFC?
>
>           
>
>         Hi, Adrian,
>
>         Thanks for raising this question. Actually vnfpool folks have been discussing this
>
>         question a lot since the very beginning of this work.
>
>         We believe vnfpool and SFC are independent and complementary mainly due to
>
>         the below reasons:
>
>         1) SFC targets on steering packets among service function nodes. vnfpool
>
>         focuses on redundancy for service nodes, e.g., selecting standby nodes,
>
>         handling nodes transition/failure cases, without caring how
>
>         2) vnfpool manager in our proposal could interact with SFC control entity to: 1)
>
>         advertise redundant service nodes; 2) notify status of redundant nodes when
>
>         required; 3) receive resiliency requirements from SFC control entity (if any); and
>
>         so on.
>
>         3) vnfpool is not only used in "chained service nodes", but applicable to other
>
>         cases where service nodes are not necessarily sequentially connected.
>
>         I appreciate any further feedback and advice from you or IESG, as I believe
>
>         these feedback will greatly improve the quality of our proposed charter.
>
>         Thanks.
>
>         -Ning
>
>           
>
>             -----Original Message-----
>
>             From: vnfpool [mailto:[e-mail]] On Behalf Of Adrian Farrel
>
>             Sent: Wednesday, January 22, 2014 1:57 AM
>
>             To:[e-mail]  <mailto:[e-mail]>
>
>             Subject: [vnfpool] Overlap with SFC?
>
>               
>
>             Hi,
>
>             The IESG is looking at the BoF requests for London, and a question came upon the overlap between the proposal here and items 4 and 5 in the SFC Charter.
>
>             I think it would be valuable if you could discuss the overlap and the
>
>             interaction between the two efforts so that there is a clear view.
>
>             Thanks,
>
>             Adrian
>
>               
>
>             _______________________________________________
>
>             vnfpool mailing list
>
>             [e-mail]  <mailto:[e-mail]>
>
>             https://www.ietf.org/mailman/listinfo/vnfpool
>