Re: [Netslices] draft-qiang-netslices-gap-analysis-00 Uploaded

[Henk Birkholz <[e-mail]>]

Hello,

I have just skimmed the first part of the document.

1.) If you pull in terms defined in other drafts or (proposed) 
standards, please include the references. (random example: MDSC is 
defined in I.D-ietf-teas-actn-framework, I think).

If it is not getting too big and if it is vital to grasp the context of 
the draft, I would suggest to pull in the complete definition (and still 
reference the source, of course).

2.) Some terms I can only guess what they mean in this context, examples 
include terminal, guarantee, or the scope of the expression "individual 
parts and pieces of the overall system", which is probably something 
like every system entity or subsystem that is required to provide the 
capabilities of a differentiable network slice?

3.) In general, I like section 4.1 very much. It enables a good 
understanding of what are important characteristics of a network slice 
that are potential KPI.

4.) I would like to see an explicit differentiation and definition of 
the terms action, task, procedure (and maybe even a "middle-layer" term 
between action and procedure called activity - a building block of 
actions that can compose procedures). The terms mentioned above imply a 
specialization (and there for a some sort of taxonomy) that can be used 
to map actual running solutions and actions that can be performed by 
them - to more general procedures that are required to create a network 
slice on an architectural level.

5.) I like the Topology Hierarchy (Stack) in Figure 2, but I am not sure 
if the expressiveness of, e.g. draft-ietf-i2rs-yang-l2-network-topology 
satisfies the requirements of network slicing. For example, top-of-the 
rack data plane units, which physical ports are mapped to multiple 
redundant endpoints that maintain a partially merged control plane (to 
handle the redundant mapping) but are separate information systems on 
the management plane cannot be modeled, I think. This is just one of the 
more complex examples that fall under the scope of resilience (see below).

6.) In general, the topic Guaranteed Slice Performance and Isolation 
does not seem to address availability or resilience of a or inside a 
network slice. Is that out of scope? That is why I would like a better 
understanding of what can be "guaranteed" in general (see item 2. above).

7.) I find the title 7.3. Abstraction of Network in Network confusing 
after reading the content of that section.


This was just a quick first pass. In general, this is an excellent 
contribution and really focuses an essential part of the ongoing 
activities! Thank you!


Viele Grüße,

Henk








On 06/02/2017 11:36 AM, qiangli (D) wrote:
> Dear All,
> 
> We have upload "Gap Analysis for Network Slicing" draft 00 on 
> https://datatracker.ietf.org/doc/draft-qiang-netslices-gap-analysis/?include_text=1
> 
> Your comments and reviews are greatly appreciated.
> 
> Abstract:
> 
>     This document presents network slicing differentiation from the non-
> 
>     partition network or from simply partition of connectivity resources.
> 
>     It lists 15 standardization gaps related to 6 key requirements for
> 
>     network slicing.  It also presents an analysis of existing related
> 
>     work and other potential solutions on network slicing.
> 
>     This gap analysis document aims to provide a basis for future works
> 
>     in network slicing.
> 
> Best regards,
> 
> Co-authors
> 
> 
> 
> _______________________________________________
> Netslices mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/netslices
>