Needed: testing and research help for wireless backhaul and wide area
topology.
We have direct line of sight with almost the whole region (pretty
well) for a location we're in talks with and which has a lot of
bandwidth (10 phone lines and a few cable lines and only a couple in
current use) already going into it. If this location comes on board
they'll want a very few discreet antennas. What kind of antennas, with
what kind of gain and E & H planes, provides the lowest interference
and the best coverage while being discreet on a building?
We'd want an average shot of about 6km to reach up to a 500,000
(approximately 170,000 households) population with our backhaul
(assuming we can get fiber into the location, which we are exploring).
Most of the distance the signal would travel would be high above the
city, and thus not get too much household device interference. But as
we'd mostly be shooting into residential towers (that'd be nice in
marketing and topology terms) so we might run into interference at
those 'vertical device stacks'.
The question then becomes topology topology topology. With the new
units - Engenius, Bullet2, Nanosations - now working well with
RO.B.IN, how many units will we require per square km, what about
interference with these high powered and sensitive things. With these
units able to adapt antennas onto them (via the Bullet or what have
you), it would seem they could catch our signal from great distances,
or could they? The Engenius 2610 (still not quite
RO.B.IN compatible)
could certainly catch our backhaul signal from a good 6km out, likely
even without an antenna modification. Someone want to verify these
specs?
I'm wanting to order some equipment to test this out in some long
distance experiments. I have family near Metrotown, which would be
about 4.7 KM from our potential high-point building. Also right below
the potential high-point building is about a 400 meter stretch of
apartment buildings with a fairly young and lower income resident
demographic, free internet for testing access might actually fly with
such people. Because these residents aren't totally destitute poor
like the DTES, they might actually have some laptops and pdas and we
wouldn't need to go through poverty pimps who mostly eshew our
proposals b/c it's extra work(?) The backhaul we could get out of this
high-point building would be much cheaper than the $1/gb Mike has with
his friend' 10/10 connection at the DTES location. $1/gb on an open
test network would have us buried under debt very quickly. We can get
24/4 to test on for about $200 a month including 1500 gb. We can
possibly get more if technicals allow us to continue.
There are a number of questions to explore with a wireless backhaul
that crosses over all the nodes to hit pre-set bridges/cluster
gateways. Centralised bachkaul might not be a problem but mesh hasn't
been designed for centralized backhaul per se (where we get fiber into
a few high points around town and shoot it to the rest of the
network). Indeed some have experience on the
RO.B.IN platform that
having wired backhaul in each gateway per cluster (of 4-7 repeaters
around that gateway) is better than centralized backhaul. But
centralised wireless backhaul is by far the cheapest and best way
around the industry-oligopoly BS.
Some questions related to ESSID (& SSID) and dual channel operation,
where the bridges/cluster gateways pick up the backhaul signal on one
and then transmit for the local cluster repeaters around it on
another, would be interesting and maybe fruitful to explore, the
Engenius might do this? What about interference by our (and others')
signal cutting across multiple mesh clusters? A problem might be that
the Open-Mesh dashboard doesn't quite allow for this, yet. So we might
start a list of what we'd like if we were to develop our own platform
or if Open-Mesh really does become an open-source project at some
point. Please, if you're interested, try to wrap your head around
this.
Anyway, if others would check up these units specs and help us think
about all these topology questions that'd be great.
Engenius, Nanostations, Bullet (antenna adapter):
https://www.open-mesh.com/store/categories.php?category=Compatible-Solutions
It would seem we're going to be very unpopular amongst industry
competitors. Even some of the independently provided backhaul options
we thought we had are disappearing; everyone wants to put on usage
caps that make it cost impossible for us to deliver an actual network.
Even our best, last remaining option, which is supposedly progressive,
is in doubt. We need to quickly get to our own fiber and thus a fairly
centralised distribution of the backhaul over high gain antennas. Yes
MetroBridge now has caps according to Scott.
Some antennas I've been looking at as discreet home options:
The tiny Luxul 15dbi circular antenna 60 degree E & H plane.http://
www.whotspot.com/catalog/Flat-Panel-Patch-Antenna-15dBi-Circular-Polarized.htm
Something like this could go in a window maybe:
http://www.tootoo.com/d-rp17372318-WIFI_Square_Grid_Parabolic_Antenna_15dBi/
Also some Yagi antennas (with the weather covering so they look sleek)
with less wide e-planes would be nice. I haven't seen many yagis
however with E&H planes that I really liked. Maybe people more
familiar with antenna stuff than myself can weigh in on the kind of
criteria we'd be looking at in various topographical situations. There
are some commercial district district buildings I have friends in who
could put proper antenna arrays up on with full flat panels attached,
so big antennas are not totally out of the question. It might well
come down to 'cold calling' some buildings to see if they want super
cheap telecommunications in exchange for their roofs and putting in a
short splice run of fiber (like under a block).
Needed: opinions, research and testing help.