Re: [bmwg] New Version for draft Methodology for VNF Benchmarking Automation

[Vladimir Vassilev <[e-mail]>]

Hi Raphael,

On 20/10/2020 21.16, Raphael Vicente Rosa wrote:
> ...
> 7. The reference implementation realizes all the methodology (getting 
> some results to present).

I am in the process of reviewing the draft. Can you post a link to some 
results (XML instance data) of a benchmark report generated according to 
the YANG data model?

That said here are some notes I made so far:

* I am a bit uncertain about what "VNF-BD YANG path of the variable" is? 
Have you considered using "type instance-identifier;"?

        leaf path {
          type string;
          description "The VNF-BD YANG path of the
          variable.";
        }

* IMO The short description in "5.3.  VNF Benchmarking Report (VNF-BR)" 
can be improved with some examples. Appendix with complete example 
report data instance file.

* IMO rfc8345 should be used for the topology definitions that are 
significant part of the /vnf-bd:scenario - nodes, connection-points, 
links etc. [e-mail] can be augmented with 
additional context specific data.

* ETS14a,b,c,e reference links have typos and do not work. ETS14d, f  work.

* use "ietf-yang-types:date-and-time" instead of "string" for timestamps.

* use "ietf-yang-types:uuid" types instead of "string"

* Add YANG tree diagrams:

    $ pyang -f tree  --path 
/usr/share/yuma/modules/ietf:/usr/share/yuma/modules/ietf-drafts:. 
vnf-bd.yang

    module: vnf-bd
      +--rw experiments
      |  +--rw trials?   uint32
      |  +--rw tests?    uint32
     +--rw scenario
      |  +--rw nodes* [id]

...

* Drafts with YANG modules should use <CODE BEGINS> file ... <CODE ENDS> 
so that the modules are detected and validated by the IETF tools.

/Vladimir