Re: [Actn] Regarding draft-lopez-actn-vno-multidomains

Hi Dhruv,

I think Diego pretty much handled all your comments. One comment I have is concerning your comment on 

>> The operation and control/management of these multiple networks
>> as a single virtualized network.
>> Since there will be multiple virtual networks, can we change to 'a single virtualized network for a customer' instead?

I think the use-case discussed in 'draft-lopez-actn-vno-multidomains' deals with having many different domains under its control and addresses the need for a global operational view on all these domains. In such context, a single virtualized network is for the operator and there is no customer involved here per se. Each domain would provide some abstraction to the VNO-coordinator in which to create a global network abstraction.

Let me know if this makes sense. 
Thanks,

Young 

-----Original Message-----
From: ACTN [mailto:actn-bounces@ietf.org] On Behalf Of Diego R. Lopez
Sent: Thursday, June 19, 2014 11:13 PM
To: Dhruv Dhody
Cc: draft-lopez-actn-vno-multidomains@tools.ietf.org; actn@ietf.org
Subject: Re: [Actn] Regarding draft-lopez-actn-vno-multidomains

Hi Dhruv,

Thanks for your comments. We'll try to address them in the document as soon as we can. A few replies inlined below...

On 18 Jun 2014, at 06:18 , Dhruv Dhody <dhruv.dhody@huawei.com> wrote:
> Can we elaborate a bit more regarding the impact of NFV? Maybe mention 
> especially virtualization of CE and PE devices? Coexistence of 
> virtualized and non-virtualized network functions etc?

I'd prefer not to go for specifically mentioning concrete cases unless you see a special aspect to be remarked (if you do, please let's talk about them). OTOH, you completely right: the coexistence or virtualized and non-virtualized functions is a key aspect, and ACTN should definitely be of help for making it easier.

>>    . The operation and control/management of these multiple networks
>>       as a single virtualized network.
>>
>
> Since there will be multiple virtual networks, can we change to 'a 
> single virtualized network for a customer' instead?

Shall que use "per tenant", following the common virtualization jargon?

>>  Figure 2 shows operations hierarchy based on Figure 1. The two main  
>> ideas are:
>>
>>  1. Domain control/management entities (e.g., DCN Domain Control A, B,
>>    C and TN Domain Control) are kept intact to continue its domain
>>    operations with its technology choice and policy, etc. As
>>    discussed before domain control/management entities can be a form
>>    of various types (e.g., SDN-controller, NMS/EMS, Control Plane, or
>>    a combination of these entities, etc.) that is responsible for
>>    domain-specific network operations.
>>
>>  2. The VNO Coordinator establishes a standard-based API (which is
>>    termed as the Virtual Network Operations Interface (VNO-I) in
>>    Figure 2) with each of the domain control/management entities. The
>>    VNO coordination takes place via the VNO-I's.
>
> Can this be aligned to the framework document and clarify that VNO 
> Coordinator is VNC while DCN and TN control are PNC?  VNO-I is VNC-PNC 
> Interface (VPI)?

This alignment you request makes a lot of sense to me. In fact, wherever there are one-to-one mappings we should use the framework document terminology

> Also I was wondering if there exist another entity customer similar to 
> the framework document?

I under the impression the customer entity is implicit here, and therefore not mentioned...

>>       It is the responsibility of domain control/management entity to
>>       create an abstraction of its network topology. The level of
>>       abstraction varies from one domain to another, subject to local
>>       domain policy. All EPs and gateway nodes to other domains need
>>       to be represented at a minimum. The level of internal nodes and
>>       links may be abstracted according to its domain policy.
>
> Do you think it is possible that domain-level control provides raw 
> topology and VNO coordinator is responsible for global abstraction?

That's an option, that implies full trust between every domain control entity and the VNO, but other degrees of trust can exist, where a given domain control could not want to share the full detail of its underlying topology to a VNO.

Be goode,

--
"Esta vez no fallaremos, Doctor Infierno"

Dr Diego R. Lopez
Telefonica I+D
http://people.tid.es/diego.lopez/

e-mail: diego@tid.es
Tel:    +34 913 129 041
Mobile: +34 682 051 091
-----------------------------------------

________________________________

Este mensaje se dirige exclusivamente a su destinatario. Puede consultar nuestra política de envío y recepción de correo electrónico en el enlace situado más abajo.
This message is intended exclusively for its addressee. We only send and receive email on the basis of the terms set out at:
http://www.tid.es/ES/PAGINAS/disclaimer.aspx
_______________________________________________
ACTN mailing list
ACTN@ietf.org
https://www.ietf.org/mailman/listinfo/actn