Muni fiber: L1 or L2?

Sat Feb 2 18:09:03 UTC 2013

Perhaps I missed a reference to receiver sensitivity in this thread. Since
the receiver optical-electric components are binary in nature, received
optical dB only has to be equal to or greater than the receiver's
sensitivity. Low or high dB received light produces the same quality at the
receiver. Thus, dB loss can be extensive due to factors such as
attenuation, splices, dispersal, but as long as the received dB level is
equal to the receiver sensitivity, it doesn't matter how much launched dB
is lost. Is the point that splitters reduce the effective distance from the
launch point in the PON architecture?

David

On Fri, Feb 1, 2013 at 7:52 PM, Owen DeLong <owen at delong.com> wrote:

>
> On Feb 1, 2013, at 14:17 , Jean-Francois Mezei <
> jfmezei_nanog at vaxination.ca> wrote:
>
> > On 13-02-01 16:03, Jason Baugher wrote:
> >
> >> The reason to push splitters towards the customer end is financial, not
> >> technical.
> >
> > It also has to do with existing fibre infrastructure. If a Telco has
> > already adopted a "fibre to a node" philosophy, then it has a;ready
> > installed a limited number of strands between CO and many neighbouhoods.
>
> Since the discussion here is about muni fiber capabilities and ideal
> greenfield
> plant designs, existing fiber is irrelevant to the discussion at hand.
>
> > It makes sense to standardise on one technology. And if that technology,
> > because it is used by many, ends up much cheaper due to economies of
> > scale, it makes sense to adopt it.
>
> Only if you're a single vendor looking to provide a single-vendor solution.
> That's really not what this conversation is about, IMHO. In fact, that's a
> pretty good summary of the situation we're trying to fix.
>
> > And remember that it isn't just the cable. You need to consider the OLT
> > cards. An OLT card can often support a few GPON systems each passing 32
> > homes.
>
> Not sure why this matters...
>
> > With 1 strand per home, you take up one port per home served. (possibly
> > per home passed depending on deployment philosophy). So you end up
> > needing far more cards in an OLT to serve the same number of people.
> > More $$$ needed.
>
> Uh, no... That's not what we're talking about. We're talking about still
> using
> splitters, but, putting the splitter next to the OLT instead of near the
> ONT
> end. That's all.
>
> > GPON isn't suited for trunks. But for last mile, is it really so bad ?
>
> Yes... Because...
>
> > 2.mumble gpbs of capacity for 32 homes yields 62mbps of sustained
> > download for each home. (assuming you have 32 homes conected and using
> > it at same time)
>
> Great by todays standards, but likely to be obsoleted within 10 years.
> Given
> the nearly 100 year old nature of some copper plants, I'd like to see us
> start
> building fiber plants in a way that doesn't lock us into a particular
> technology
> choice constrained to the economic tradeoffs that are relevant today and
> may be completely different in as little as 5 years.
>
> > If you have multicast and everyone is watching superbowl at same time,
> > you're talking up very little bandwidth on that 2.mumble GPON link.
>
> Meh. Since everyone seems to want to be able to pause, rewind, etc.,
> multicast doesn't tend to happen so much even in the IPTV world these
> days.
>
> Owen
>
>
>
>