Re: [manet] Alissa Cooper's No Objection on draft-ietf-manet-dlep-26: (with COMMENT)

[Rick Taylor <[e-mail]>]

Hi Alissa,

On Thu, 2016-12-15 at 09:09 -0500, Alissa Cooper wrote:
> Hi Rick,
> 
> > 
> > On Dec 15, 2016, at 5:35 AM, Taylor, Rick (External) <rick.taylor.e
> > [e-mail]> wrote:
> > 
> > Hi Alissa,
> > 
> > Thank you for the review.  A few replies to your comments inline...
> > 
> > (Sorry for the multiple postings, but my local mailserver hates the
> > IETF mailing lists)
> > 
> > > 
> > > -----Original Message-----
> > > From: Alissa Cooper [mailto:[e-mail]]
> > > Sent: 14 December 2016 19:06
> > > To: The IESG
> > > Cc: [e-mail]; [e-mail]; manet-
> > > [e-mail]; [e-mail]; [e-mail]
> > > Subject: Alissa Cooper's No Objection on draft-ietf-manet-dlep-
> > > 26:
> > > (with
> > > COMMENT)
> > > 
> > > Alissa Cooper has entered the following ballot position for
> > > draft-ietf-manet-dlep-26: No Objection
> > > 
> > > When responding, please keep the subject line intact and reply to
> > > all
> > > email addresses included in the To and CC lines. (Feel free to
> > > cut
> > > this introductory paragraph, however.)
> > > 
> > > 
> > > Please refer to
> > > https://www.ietf.org/iesg/statement/discuss-criteria.html
> > > for more information about IESG DISCUSS and COMMENT positions.
> > > 
> > > 
> > > The document, along with other ballot positions, can be found
> > > here:
> > > https://datatracker.ietf.org/doc/draft-ietf-manet-dlep/
> > > 
> > > 
> > > 
> > > ---------------------------------------------------------------
> > > -------
> > > COMMENT:
> > > ---------------------------------------------------------------
> > > -------
> > > 
> > > For the Latency and Relative Link Quality metrics, it seems like
> > > allowing their measurement methods to be implementation-dependent
> > > reduces their usefulness. If two different implementations
> > > calculate
> > > these in different ways, then the results may not be comparable.
> > > Are
> > > these not meant to be compared between different destination
> > > implementations?
> > Yes, this is indeed a difficult issue.  The underlying concept
> > behind DLEP is that modem implementers have their own low-level
> > mechanisms for measuring such things as latency, and exposing these
> > metrics to the routing layer rather than forcing the up layers to
> > measure such things can be useful.  However, we didn't want to
> > prescribe how a lower-layer might calculate such metrics, as we
> > couldn't see a 'one size fits all' mechanism being useful.
> > 
> > When comparing metrics between destinations reported by a single
> > modem, it is considered a safe assumption that the metrics are
> > sensibly comparable.  When comparing metrics reported between
> > different modems, it gets more complicated.  For metrics such as
> > latency, it's fairly well understood what latency means.  
> I believe that, but as Mirja points out, the document doesn’t specify
> whether what is being reported is an average or max or min over some
> period of time. It seems like at least a little more detail could be
> provided to make this one more useful. 

Ironically, there is a draft that has just been adopted by the WG, that
describes a DLEP extension for min/max for the latency metric.

The first paragraph of Section 4 attempts to suggest that metrics
should be an (perhaps rolling) average over a "...large enough sample
sizes will preclude short traffic bursts from adversely skewing
reported values..."

But perhaps that text needs improving? 

Thanks,

Rick