Verizon Policy Statement on Net Neutrality

Daniel Taylor dtaylor at vocalabs.com
Fri Feb 27 21:05:24 UTC 2015


On 02/27/2015 02:53 PM, Scott Helms wrote:
> "My point is that the option should be there, at the consumer level."
>
> Why?  What's magical about symmetry?  Is a customer better served by 
> having a 5mbps/5mbps over a 25mbps/5mbps?
>
Why not 25/25?

50MB/s might be tough to fill, but even at home I can get good use out 
of the odd 25MB/s upstream burst for a few minutes.

>
> "There are so many use cases for this, everything from personal game 
> servers to on-line backups, that the lack of such offerings is an 
> indication of an unhealthy market."
>
> Until we get NAT out of the way, this is actually much harder to 
> leverage than you might think.  I don't think there is anything 
> special about symmetrical bandwidth, I do think upstream bandwidth 
> usage is going up and will continue to go up, but I don't see any 
> evidence in actual performance stats or customers sentiment to show 
> that it's going up as fast as downstream demand.
>
>
> Scott Helms
> Vice President of Technology
> ZCorum
> (678) 507-5000
> --------------------------------
> http://twitter.com/kscotthelms
> --------------------------------
>
> On Fri, Feb 27, 2015 at 3:36 PM, Daniel Taylor <dtaylor at vocalabs.com 
> <mailto:dtaylor at vocalabs.com>> wrote:
>
>     My point is that the option should be there, at the consumer level.
>
>     If not for fully symmetrical service (I admit that 50MB/s upstream
>     is a tough pipe to fill), at least for significantly higher
>     upstream service than is currently available in most neighborhoods.
>
>     There are so many use cases for this, everything from personal
>     game servers to on-line backups, that the lack of such offerings
>     is an indication of an unhealthy market.
>
>     On 02/27/2015 02:25 PM, Scott Helms wrote:
>
>         Daniel,
>
>         We'd have to come to some standard definition of, "But even if
>         1% of users would reasonably be using a fully symmetric link
>         to its potential..."
>
>         As I said, I have visibility into a large number of symmetric
>         connections and without exception they'd fit well into a plan
>         that offered upstreams with that had a fractional speed of the
>         downstream.  Now, keep in mind I'm not talking about 1/10 as a
>         ratio here, but 1/5 would accommodate ~99.2% and 1/4 would fit
>         ~99.9%.  It's also important to note that all of these
>         accounts are in the >25mbps down territory so their upstreams
>         are >5mbps.
>
>         What I see when I look at customer satisfaction ratings is a
>         very strong correlation with low uplink speeds and a high
>         satisfaction rate when we look at uplink speeds greater than
>         4mbps.  What I don't see is an increase in customer
>         satisfaction as upload speeds go past ~6mbps. Conversely,
>         increases in customer satisfaction with correlate with
>         increases in download speeds past ~30mbps before the
>         correlation starts weakening.
>
>
>         Scott Helms
>         Vice President of Technology
>         ZCorum
>         (678) 507-5000 <tel:%28678%29%20507-5000>
>         --------------------------------
>         http://twitter.com/kscotthelms
>         --------------------------------
>
>         On Fri, Feb 27, 2015 at 2:57 PM, Daniel Taylor
>         <dtaylor at vocalabs.com <mailto:dtaylor at vocalabs.com>
>         <mailto:dtaylor at vocalabs.com <mailto:dtaylor at vocalabs.com>>>
>         wrote:
>
>             The statistics certainly *should* be used when provisioning
>             aggregate resources.
>             But even if 1% of users would reasonably be using a fully
>             symmetric link to its potential, that's a good reason to
>         at least
>             have such circuits available in the standard consumer mix,
>         which
>             they aren't today.
>
>             On 02/27/2015 01:30 PM, Scott Helms wrote:
>
>                 Daniel,
>
>                 Well, I wouldn't call using the mean a "myth", after all
>                 understanding most customer behavior is what we all
>         have to
>                 build our business cases around.  If we throw out what
>                 customers use today and simply take a build it and
>         they will
>                 come approach then I suspect there would fewer of us
>         in this
>                 business.
>
>                 Even when we look at anomalous users we don't see
>         symmetrical
>                 usage, ie top 10% of uploaders.  We also see less
>         contended
>                 seconds on their upstream than we do on the
>         downstream.  These
>                 observations are based on ~500k residential and business
>                 subscribers across North America using FTTH (mostly GPON),
>                 DOCSIS cable modems, and various flavors of DSL.
>
>
>                 Scott Helms
>                 Vice President of Technology
>                 ZCorum
>         (678) 507-5000 <tel:%28678%29%20507-5000>
>         <tel:%28678%29%20507-5000>
>                 --------------------------------
>         http://twitter.com/kscotthelms
>                 --------------------------------
>
>                 On Fri, Feb 27, 2015 at 2:21 PM, Daniel Taylor
>                 <dtaylor at vocalabs.com <mailto:dtaylor at vocalabs.com>
>         <mailto:dtaylor at vocalabs.com <mailto:dtaylor at vocalabs.com>>
>                 <mailto:dtaylor at vocalabs.com
>         <mailto:dtaylor at vocalabs.com> <mailto:dtaylor at vocalabs.com
>         <mailto:dtaylor at vocalabs.com>>>>
>                 wrote:
>
>                     But by this you are buying into the myth of the mean.
>
>                     It isn't that most, or even many, people would take
>                 advantage of
>                     equal upstream bandwidth, but that the few who
>         would need
>                 to take
>                     extra measures unrelated to the generation of that
>         content
>                 to be
>                     able to do so.
>
>                     Given symmetrical provisioning, no extra measures
>         need to
>                 be taken
>                     when that 10 year old down the street turns out to
>         be a master
>                     musician.
>
>                     On 02/27/2015 11:59 AM, Scott Helms wrote:
>
>                         This is true in our measurements today, even when
>                 subscribers
>                         are given
>                         symmetrical connections.  It might change at some
>                 point in the
>                         future,
>                         especially when widespread IPv6 lets us get
>         rid of NAT
>                 as a de
>                         facto
>                         deployment reality.
>
>
>                         Scott Helms
>                         Vice President of Technology
>                         ZCorum
>         (678) 507-5000 <tel:%28678%29%20507-5000>
>         <tel:%28678%29%20507-5000>
>                 <tel:%28678%29%20507-5000>
>                         --------------------------------
>         http://twitter.com/kscotthelms
>                         --------------------------------
>
>                         On Fri, Feb 27, 2015 at 12:48 PM, Naslund, Steve
>
>
>     -- 
>     Daniel Taylor          VP Operations            Vocal
>     Laboratories, Inc.
>     dtaylor at vocalabs.com <mailto:dtaylor at vocalabs.com>
>     http://www.vocalabs.com/ (612)235-5711 <tel:%28612%29235-5711>
>
>


-- 
Daniel Taylor          VP Operations            Vocal Laboratories, Inc.
dtaylor at vocalabs.com   http://www.vocalabs.com/            (612)235-5711




More information about the NANOG mailing list