Re: WG ADOPTION: https://tools.ietf.org/html/draft-vsperf-bmwg-vswitch-opnfv

Hello Al and Maryam,

Please see my comments inline.

On 4/2/2016 1:46 AM, MORTON, ALFRED C (AL) wrote:
>
> Hi Marius,
>
> Thanks for your support on the draft.
>
> Maryam and I discussed your comments briefly at the
>
> VSPERF weekly meeting.
>
> When it comes to comparison, which may not necessarily
>
> be degradation, (I think) we’ve seen multipliers used
>
> more frequently, as they help to simplify the result
>
> (2x better performance, AAA switch 10x faster than ZZZ).
>
> But this is worth discussing.
>
I think putting  a number on the performance difference between physical 
and virtual switches (which, I agree, may not necessarily be 
degradation) would be beneficial. Multipliers sound like a good solution 
as well.

> As far as De-activation goes, VSPERF doesn’t have test
>
> cases that cover this function, and it’s probably a
>
> topic which has seen light treatment in Benchmarking.
>
> But, we could turn that around if it’s important now!
>
I don't think it's terribly important. I was just wondering why, since 
it's easily noticeable from the cover matrix.

Best regards,
Marius

> *From:*bmwg [mailto:[email protected]] *On Behalf Of *Marius Georgescu
> *Sent:* Tuesday, March 29, 2016 11:07 PM
> *To:* [email protected]
> *Subject:* Re: [bmwg] WG ADOPTION: 
> https://tools.ietf.org/html/draft-vsperf-bmwg-vswitch-opnfv
>
> I think this work is important for the future of this WG and I support 
> its adoption.
>
> I also have two questions for the authors:
>
> 1. For the Comparison with Physical Network Functions, I was wondering 
> if a performance degradation index between a physical and a virtual 
> network function would make sense. For example, run a 
> RFC2544.throughput test with a physical switch as DUT and  get X 
> result, run the same test with a  virtual switch as DUT and get Y 
> result. Express the performance degradation in percentage as:
>
>               X - Y
>
>        PD= ----------- * 100
>
>                  X
>
> 2. Considering the Matrix Coverage Summary, I was wondering why no 
> De-activation benchmark was proposed.
>
> Best regards,
>
> Marius
>
>     On Mar 24, 2016, at 00:46, Sarah B <[email protected]
>     <mailto:[email protected]>> wrote:
>
>     Hello BWWG,
>
>     According to the minutes from IETF94
>     (https://www.ietf.org/proceedings/94/minutes/minutes-94-bmwg),
>     those present in the room expressed interest in the draft.
>     There've been several reviewers, and strong discussion during each
>     IETF meeting since the draft was introduced. As Al is a named
>     author, I'd like to, with my chair hat on, ask the BMWG list to
>     confirm their support for this draft and call for adoption.
>
>     Please reply with your opinion whether the BMWG should adopt this
>     draft, whether you have read the draft, and any comments on the
>     draft or this call for adoption. Let's attempt to complete this
>     call for adoption and meet a deadline of April 20th, 2016.
>
>     Kind regards,
>
>     Sarah
>
>     BMWG Co-Chair
>
>     _______________________________________________
>     bmwg mailing list
>     [email protected] <mailto:[email protected]>
>     https://www.ietf.org/mailman/listinfo/bmwg
>

_______________________________________________
bmwg mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/bmwg