Rollup: Small City Municipal Broadband

Scott Helms khelms at zcorum.com
Sun Feb 3 22:14:54 UTC 2013


On Sun, Feb 3, 2013 at 4:58 PM, Jay Ashworth <jra at baylink.com> wrote:

> ----- Original Message -----
> > From: "Scott Helms" <khelms at zcorum.com>
>
> > > And flooding doesn't affect pure glass, does it?
> >
> > Not directly, so long as the cladding stays intact. The problem with
> > flooding (for your scenario since your electronics will be centralized)
> is
> > mainly that it causes things to move around inside the cable runs and
> > depending on water flow you can end up with a lot of problems with
> > increased scattering if the cable gets stretched.
>
> Yeah.  The cable has like 2 or 3 more layers, including a strength
> member, above the cladding, though, right?
>
Yep, but glass is much (an order of magnitude) more sensitive to stretching
than coax or twisted pair (electrons don't mind, but light diffracts).
 That's not to say that aerial isn't doable it is, but you're going to be
doing maintenance on your plant most years.

>
> > > It would appear that opinions vary on this point. You've clearly
> > > had your hands on some of the gear, so I'm not discounting your opinion
> > > by any means, but it seems that this may vary based on, among other
> > > things, how well one engineers the plant up front. This will *not*
> > > be a lowest-bidder contract. Or I won't do it.
> >
> > Its not just the initial install. Its that every time you do anything new
> > like adding in a new L2 vendor or technology or hook up a new end user.
>
> Since I plan to drop and terminal all the tails on the initial install,
> "hook up a new user" amounts to "plug an optical patch cord into a wall
> jack".
>
> Assuming the termination of the fiber in the wall jack was tested for
> level and blur at install time, that will hopefully not be too big
> a deal.
>

How is it tested?  Just looking at it with an OTDR meter doesn't mean it
will work with RFoG or PON.

>
>

> > > As Owen notes, their hot-potatoing it will simply cost them more money,
> > > so they have incentive to be cooperative in finding these problems, and
> > > that helps almost an order of magnitude.
> >
> > Respectfully the guys that will be doing the hot potato shuffle won't
> > be the owners or even people who have that much technical understanding.
>
> Well, in my particular instance, I don't actually think that will be true;
> I expect to be dealing with either the guy who cuts my check, or the
> technical lead who works for him, most of the time.
>
> I would be pretty surprised to find that the ISP I have in mind as my
> first mover has more than 10 employees.
>

Sure, but are those 10 guys the only ones who do end user installs?  If so
then you will hopefully have fewer problems, but that depends on how good
those 10 people are and the kinds of technologies you're working on.

>
> > They'll be installers that work on the end user and their skill set is
> > on par with the guys who do contract installs for security systems and
> > Dish Network/Direct TV. They don't care if their boss loses money, they
> > don't care if you lose money, all they want to is keep their install
> count
> > up since that's how they get paid. If a given install is problematic and
> > they can shift the responsibility to someone else they (as a group) will.
> > I'm not suggesting that everyone who does that kind of work is unskilled
> > or uncaring but as a group that's what you get.
>
> In general, I think that's true.  In a city of 11000 people, I am not
> sure that I think it will actually work out that way in practice.
>


I'd disagree, that's definitely of the scale where you'll have 4-5 people
who do this kind of contract installs on a regular basis.  Now, your ISP
may choose to not use one of them, but they're certainly out there.


>
> > > No, but I'm pretty sure my Fluke rep will be happy to sell me boxes
> that
> > > *will* test for that stuff, and I will have a contractor to back me up.
> >
> > No, actually they won't because Fluke doesn't sell a DOCSIS analyzer
> > (for RFoG) nor a PON analyzer. You'll need a separate meter (for several
> > thousand dollars) for each kind of technology you want to be able to
> > troubleshoot. For example, to handle modulated RF (RFoG) you'd use a
> > JDSU (or Sunrise or Trilithic). Fluke is a very basic OTDR tool and they
> > don't address the various layer 2 technologies.
>
> Well, my snap reaction to this is "what the city's responsibility is
> for dark fiber pairs is spelled out in the contract, and things past
> that test regime are the responsibility of the L2 provider", but I
> gather you don't think that will be good enough.
>
> Why is protocol my responsibility?
>

> *It's a dark pair of glass fibers*.  If it meets level and dispersion
> specs, how the hell is it still my problem?
>

That's what I've been trying to tell you.  You're basing your idea of
troubleshooting on AM or FM modulation when all of the broadband
connectivity is PM (phase modulation).  That means your light power meter
from Fluke can tell you everything is good when a specific connection is in
fact unusable.

http://en.wikipedia.org/wiki/Phase_modulation

Take a look at the write up a friend of mine did on a QAM analyzer.  This
one is specifically for cable systems, but the concept of phase
constellations works the same in any PM set up.

http://volpefirm.com/tech-review-devisor-s7000-tv-analyzer/

This PDF (search for constellation) describes how phase modulation is used
in  EPON:

http://www.broadcom.com/docs/features/DOCSIS_EoC_EPONinChina.pdf




> Cheers,
> -- jra
> --
> Jay R. Ashworth                  Baylink
> jra at baylink.com
> Designer                     The Things I Think                       RFC
> 2100
> Ashworth & Associates     http://baylink.pitas.com         2000 Land
> Rover DII
> St Petersburg FL USA               #natog                      +1 727 647
> 1274
>
>


-- 
Scott Helms
Vice President of Technology
ZCorum
(678) 507-5000
--------------------------------
http://twitter.com/kscotthelms
--------------------------------



More information about the NANOG mailing list