Muni Fiber (was: Re: last mile, regulatory incentives, etc)

Anurag Bhatia me at anuragbhatia.com
Thu Mar 29 18:25:25 UTC 2012


Thanks Jacob and Alex.


Appreciate your reply.

On Thu, Mar 29, 2012 at 8:39 AM, Jacob Broussard <
shadowedstrangerlists at gmail.com> wrote:

> While I can't provide an average, I can say we generally have anywhere
> from 2-5 microwaves on most sites (with a few exceptions that only have 1,
> and a few that have more.)  Our MWs go up to 1.6gbps.  The sites aren't
> provisioned a set amount of bandwidth, they can use as much as they want
> (up to the capacity of the aggregate of their links), which almost never
> puts our BH anywhere near capacity, unless the ring gets cut near the pop
> and we have to move lots of data through just a couple of sites. (Sorry for
> the crappy formatting, small and barely usable phone screen.)
>
> Thanks!
> -Jacob
> On Mar 28, 2012 1:45 AM, "Anurag Bhatia" <me at anuragbhatia.com> wrote:
>
>> Hi
>>
>> Nice discussion. Just a small question here - how much backhaul  at
>> present
>> 2G, 3G and LTE based towers have? Just curious to hear an average number.
>> I
>> agree it would be  a significant difference from busy street in New York
>> to
>> less crowded area say in Michigan but what sort of bandwidth telcos
>> provision per tower?
>>
>> On fiber - I can imagine virtually unlimited bandwidth with incremental
>> cost of optical instruments but how much to wireless backhaul based sites?
>> Do they put Gigabit microwave everywhere?
>>
>> If not then say 100Mbps? If so then how end users on Verizon LTE people
>> individual users get 10Mbps and so on? Is that operated at high
>> contention?
>>
>> Thanks!
>>
>> (Sent from my mobile device)
>>
>> Anurag Bhatia
>> http://anuragbhatia.com
>> On Mar 27, 2012 10:26 PM, "Alexander Harrowell" <a.harrowell at gmail.com>
>> wrote:
>>
>> > On Tue, Mar 27, 2012 at 1:45 AM, William Herrin <bill at herrin.us> wrote:
>> >
>> > > On Mon, Mar 26, 2012 at 8:04 PM, Jacob Broussard
>> > > <shadowedstrangerlists at gmail.com> wrote:
>> > > > Who knows what technology will be like in 5-10 years?  That's the
>> whole
>> > > > point of what he was trying to say.  Maybe wireless carriers will
>> use
>> > > > visible wavelength lasers to recievers on top of customer's houses
>> for
>> > > all
>> > > > we know.  10 years is a LONG time for tech, and anything can happen.
>> > >
>> > >
>> > Regarding lasers. I agree that modulating a laser beam to carry
>> information
>> > is a great idea. Perhaps, though, we could direct the beam down some
>> sort
>> > of optical pipe or waveguide to spare ourselves the refractive losses
>> and
>> > keep the pigeons and rain and whatnot out of the Fresnel zone. We might
>> > call it an "optical wire" or "optical fibre" or something. no, it'll
>> never
>> > catch on...
>> >
>> > Hi Jacob,
>> > >
>> > > The scientists doing the basic research now know. It's referred to as
>> > > the "technology pipeline." When someone says, "that's in the pipeline"
>> > > they mean that the basic science has been discovered to make something
>> > > possible and now engineers are in the process of figuring out how to
>> > > make it _viable_. The pipeline tends to be 5 to 10 years long, so
>> > > basic science researchers are making the discoveries *now* which will
>> > > be reflected in deployed technologies 10 years from now.
>> > >
>> >
>> >
>> > I recall an Agilent Technologies presentation from a couple of years
>> back
>> > that demonstrated that historically, the great majority of incremental
>> > capacity on cellular networks was accounted for by cell subdivision.
>> Better
>> > air interfaces help, more spectrum helps, but as the maximum system
>> > throughput is roughly defined by (spectral efficiency * spectrum)*
>> number
>> > of cells (assuming an even traffic distribution and no intercell
>> > interference or re-use overhead, for the sake of a finger exercise),
>> > nothing beats more cells.
>> >
>> >
>> > As a result, the Wireless Pony will only save you if you can find a
>> 10GigE
>> > Backhaul Pony to service the extra cells. After a certain degree of
>> > density, you'd need almost as much fibre (and more to the point, trench
>> > mileage) to service a couple of small cells per street as you would to
>> > *pass the houses in the street with fibre*.
>> >
>> >
>> > One of the great things FTTH gets you is a really awesome backhaul
>> network
>> > for us cell heads. One of the reasons we were able to roll out 3G in the
>> > first place was that DSL got deployed and you could provision on two or
>> a
>> > dozen DSL lines for a cell site.
>> >
>> >
>> > You can't have wireless without backhaul (barring implausible
>> discoveries
>> > in fundamental mesh network theory). Most wireless capacity comes from
>> cell
>> > subdivision. Subdivision demands more backhaul.
>> >
>> >
>> > > There is *nothing* promising in the pipeline for wireless tech that
>> > > has any real chance of leading to a wide scale replacement for fiber
>> > > optic cable. *Nothing.* Which means that in 10 years, wireless will be
>> > > better, faster and cheaper but it won't have made significant inroads
>> > > replacing fiber to the home and business.
>> > >
>> > > 20 years is a long time. 10 years, not so much. Even for the long
>> > > times, we can find the future by examining the past. The duration of
>> > > use of the predecessor technology (twisted pair) was about 50 years
>> > > ubiquitously deployed to homes. From that we can make an educated
>> > > guess about the current one (fiber). Fiber to the home started about
>> > > 10 years ago leaving about 40 more before something better might
>> > > replace it.
>> > >
>> > > Regards,
>> > > Bill Herrin
>> > >
>> > >
>> > >
>> > > --
>> > > William D. Herrin ................ herrin at dirtside.com
>> bill at herrin.us
>> > > 3005 Crane Dr. ...................... Web: <http://bill.herrin.us/>
>> > > Falls Church, VA 22042-3004
>> > >
>> > >
>> >
>>
>


-- 

Anurag Bhatia
anuragbhatia.com
or simply - http://[2600:3c01:e000:1::5] if you are on IPv6 connected
network!

Twitter: @anurag_bhatia <https://twitter.com/#!/anurag_bhatia>
Linkedin: http://linkedin.anuragbhatia.com



More information about the NANOG mailing list