What to use for long links, 30-50km ?
Valent Turkovic
Song, Stephen
Hi, we have used Mikrotik equipment previously for long range links, but as we are transitioning all to OpenWrt supported equipment I'm wondering who has experience with equipment that supports OpenWrt and has great performance on long range links.Has anybody compared Ubiquiti and Mikrotik equipment on 40-50km links?
--Cheers from Croatia,Valent.
You received this message because you are subscribed to the Google Groups "Village Telco Development Community" group.
To unsubscribe from this group and stop receiving emails from it, send an email to village-telco-...@googlegroups.com.
To post to this group, send email to village-...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/village-telco-dev/be222510-78aa-47d4-ab1a-6ac1d7ec8f8e%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
Steve Song
Dave Duchesneau
Hi Valent,
> … long links where a TDMA-based solution like Ubiquiti Airmax … will be more efficient?
Although Ubiquiti gear can also be flashed with OpenWRT, but you won’t want to do that for long links, primarily due to acknowledgment (ACK) timeout issues. If the actual link distance exceeds hardware ACK timeout limits, then various auto-adjust mechanisms fail. With 40 MHz channels, the ACK timeout limits occur at about 27 km (17 miles). Narrower channels extend this range at the expense of maximum theoretical data rate, but wider channels are useless anyway if they don’t connect or are unreliable. Narrowing the channel from 40 MHz to 20 MHz cuts the data rate in half, but extends the ACK timeout limits from 27 km out to 51 km (32 miles). With Ubiquiti gear, you can cut the channel width still further, such as to 10 MHz or 5 MHz channels (with a halving of the max data rate each time you halve the channel width).
Note that extending the ACK timeout limit from 27 km to 51 km does NOT ensure that you can communicate reliably at that distance, but it does mean that speed-of-light issues won’t prevent ACKs from being received in a timely manner. All the usual LOS issues remain (path loss, obstructions, Fresnel zone, etc.). This is where narrower channels can help a lot, since halving the channel width doubles the spectral power density (the same amount of transmit power is concentrated into a smaller channel width).
Using MIMO gear such as Ubiquiti’s Rocket M series can help restore some of that data rate loss. Since the Rocket M series is 2x2 MIMO, the data rate is doubled over that of standard 802.11n node having only a single stream. In my opinion, one of the lowest-cost routes to achieve reliable long range operation is to take advantage of the 2x2 MIMO available on the Ubiquiti Rocket M (on each end), along with the TDMA protocol in the built-in airMAX OS, combined with a high transmit power (28 dBm or 630 mW) as needed, narrow channel widths, and high-gain directional antennas. As noted earlier, you still have to address the usual LOS issues.
With regard to ACK timeouts, Ubiquiti’s airMAX OS has two different modes which are mutually exclusive. The standard mode is “Auto-Adjust” (where the ACK timeout settings are limited only by device hardware specifications, just as OpenWRT would be). The other mode is “Long Range PtP Link Mode” (where the ACK timeout settings are controlled by manually entering the distance to the client, since long links exceed the capabilities of auto-adjustment).
I hope this helps.
Dave Duchesneau
To view this discussion on the web visit https://groups.google.com/d/msgid/village-telco-dev/CAD_CWO2P7ocaDt%2BuNcSoTOnR5t9CtNEV9VVmNhrgCrUQqKwkUw%40mail.gmail.com.
paulc...@gmail.com
Paul Colmer
Valent Turkovic
Hi Valent,On 23 June 2014 11:30, Valent Turkovic <valent....@gmail.com> wrote:
Hi, we have used Mikrotik equipment previously for long range links, but as we are transitioning all to OpenWrt supported equipment I'm wondering who has experience with equipment that supports OpenWrt and has great performance on long range links.Has anybody compared Ubiquiti and Mikrotik equipment on 40-50km links?I wonder whether it makes sense to use OpenWRT with standard WiFi drivers as the CSMA protocol is not really well-designed for long links where a TDMA-based solution like Ubiquiti Airmax or Mikrotik NV2 will be more efficient?-Steve
Valent Turkovic
Hi Valent,
> … long links where a TDMA-based solution like Ubiquiti Airmax … will be more efficient?
Although Ubiquiti gear can also be flashed with OpenWRT, but you won’t want to do that for long links, primarily due to acknowledgment (ACK) timeout issues. If the actual link distance exceeds hardware ACK timeout limits, then various auto-adjust mechanisms fail. With 40 MHz channels, the ACK timeout limits occur at about 27 km (17 miles). Narrower channels extend this range at the expense of maximum theoretical data rate, but wider channels are useless anyway if they don’t connect or are unreliable. Narrowing the channel from 40 MHz to 20 MHz cuts the data rate in half, but extends the ACK timeout limits from 27 km out to 51 km (32 miles). With Ubiquiti gear, you can cut the channel width still further, such as to 10 MHz or 5 MHz channels (with a halving of the max data rate each time you halve the channel width).
Note that extending the ACK timeout limit from 27 km to 51 km does NOT ensure that you can communicate reliably at that distance, but it does mean that speed-of-light issues won’t prevent ACKs from being received in a timely manner. All the usual LOS issues remain (path loss, obstructions, Fresnel zone, etc.). This is where narrower channels can help a lot, since halving the channel width doubles the spectral power density (the same amount of transmit power is concentrated into a smaller channel width).
Using MIMO gear such as Ubiquiti’s Rocket M series can help restore some of that data rate loss. Since the Rocket M series is 2x2 MIMO, the data rate is doubled over that of standard 802.11n node having only a single stream. In my opinion, one of the lowest-cost routes to achieve reliable long range operation is to take advantage of the 2x2 MIMO available on the Ubiquiti Rocket M (on each end), along with the TDMA protocol in the built-in airMAX OS, combined with a high transmit power (28 dBm or 630 mW) as needed, narrow channel widths, and high-gain directional antennas. As noted earlier, you still have to address the usual LOS issues.
With regard to ACK timeouts, Ubiquiti’s airMAX OS has two different modes which are mutually exclusive. The standard mode is “Auto-Adjust” (where the ACK timeout settings are limited only by device hardware specifications, just as OpenWRT would be). The other mode is “Long Range PtP Link Mode” (where the ACK timeout settings are controlled by manually entering the distance to the client, since long links exceed the capabilities of auto-adjustment).
I hope this helps.
Valent Turkovic
Has anybody compared Ubiquiti M5 vs Ubiquiti Rocket for long range links?
Valent Turkovic
Is there any advantage or disadvantage to using multiple streams (mimo) to single stream equipment for long range links?
Has anybody compared Ubiquiti M5 vs Ubiquiti Rocket for long range links?
Nicolás Echániz
link here:
http://blog.altermundi.net/article/completamos-el-enlace-de-50km/
(spanish and english)
Cheers,
Nico
Dave Duchesneau
Valent, re your question on single vs. multiple MIMO streams:
> Is there any advantage or disadvantage to using multiple streams
> (mimo) to single stream equipment for long range links?
> Has anybody compared Ubiquiti M5 vs Ubiquiti Rocket for long range
> links?
MIMO streams (i.e., the Rocket M over the Bullet M)." The advantage boils
down to twice the maximum data rate at any given power level, all else being
equal. My longer answer follows.
[Note: Before I get into the longer answer, I just noticed that you
specifically mentioned Ubiquiti's M5 line, whereas I used the M2 line in my
examples below. Everything I say below applies equally to the M5 line, the
M9 line, etc. Discussing the merits of M2 vs. M5 vs. M9, and so forth is
another discussion entirely, and due to the need to address the specifics of
the links to be established. A useful online tool for rough calculations
based on topology, range, antenna height, etc., with endpoints specified via
Google Earth, can be found at: http://www.ubnt.com/airlink/ -- Dave]
POWER: Regulatory agencies limit the maximum transmitted power output from
each node. Given a maximum power 'P' and a number of streams 'S' (which
also implies 'S' radios transmitting in unison at each node), the power must
be divided across the streams, where each stream transmits at a power level
of P/S, so that the total doesn't exceed the allowed maximum. For example,
a Bullet M2 operating at 26 dBm (less than max) would deliver full power P
(~400 mW) to the one stream (P/S = 400/1 = 400). A Rocket M2 operating at
26 dBm (also less than max) would deliver the same TOTAL power P (~400 mW)
by transmitting it over two streams (P/S = 400/2 = 200), with each stream's
radio operating at 23 dBm (3dB less is a halving of power) and delivering
half the total power.
MAXIMUM DATA RATE: Each stream has a maximum data rate that it can support,
based on the number of symbols that can be communicated from one point to
another. With one stream you get whatever the stream is capable of, minus
whatever losses are incurred due to path loss, fading, and multipath, etc.
Historically, multipath is a "bad" thing, because bounced/reflected signals
arriving out of order can confuse the receiver, resulting in reduced data
rates. However, MIMO is kind of like intentional multi-path where the
multiple streams arriving at the receiver have mathematical relationships
among themselves (introduced prior to transmission). The combination of
streams carries more information (symbols), in direct relation to the number
of streams than would be present in a single stream. In our example, a
Bullet M2 is like 1x1 MIMO (1 transmitter, 1 receiver), which can handle
only one stream and is really not MIMO at all (MIMO stands for multiple
input, multiple output). A Rocket M2 implements 2x2 MIMO (2 transmitters, 2
receivers) and can deliver two streams (i.e., twice as much data) at the
same level of power. If Ubiquiti had a 4x4 model, it would deliver four
times the data rate. The downside of MIMO is the need for extra antennas
which do not interfere with each other very much. This is easy with 2x2
MIMO, since each linearly polarized antenna can be a different polarity
(horizontal or vertical). Beyond that, avoiding interference gets tricky
(e.g., left or right-hand circular polarization, etc.).
Power Spectral Density (PSD): Typical WiFi gear operates with 20 MHz
channels. Doubling the channel width doubles the maximum data rate, but
reduces the range, because its PSD gets halved (the same amount of power
gets spread more 'thinly' across a wider channel). Conversely, halving the
channel width increases the range, because the PSD gets doubled (the same
amount of power gets concentrated across a narrower channel).
MIMO stream and PSD: The relationship between MIMO streams and PSD is very
important. Increasing the PSD (e.g., by narrowing the channel width)
increases the effective range, but at the expense of data rate. The use of
multiple streams linearly increases the data rate (2 streams = 2X data rate)
at the expense of an extra radio (internal) and antenna on each end. For
example, given a Bullet 2M operating at 26 dBm (~400 mW), narrowing the
channel width from 20 MHz to 10 MHz will double the power concentrated into
the channel width (providing increased range), but will halve the maximum
data rate, with no recourse since there is only one stream. In contrast,
given a Rocket M2, the same situation will occur in terms of range increase
due to concentrated power, with one important exception. With the 2x2 MIMO,
the halving of the data rate due to narrowing the channel width will be
compensated by the doubling of the data rate due to its 2x2 MIMO streams.
The net result is that Rocket M2 can always match any range that the Bullet
M2 can do, but with a data rate that is roughly double.
TDMA vs. CSMA on Long-Range Links: Besides the improved suitability of the
TDMA protocol over CSMA on long-range links, if you're considering a
non-Ubiquiti implementation of TDMA, be sure that you can change the timing
of ACK timeouts (as discussed in my previous email) to meet your needs. The
option to manually configure ACK timeouts (based on distance) in Ubiquiti's
long-haul gear is due to their desire to provide carrier-quality equipment,
so it is not necessarily available in alternative TDMA implementations.
Also, Ubiquiti's TDMA protocol is built in as an option and does not require
any extra costs or license keys. Just something to keep in mind.
Good luck!
Dave Duchesneau
d...@crisis-force.org
P.S. In case I sound like I'm giving a sales pitch, let me include a
disclosure statement.
FULL DISCLOSURE: For the record, I don't work for Ubiquiti, don't invest in
Ubiquiti, and I have no relationships with anyone who does. I have,
however, acquired numerous Ubiquiti radios from various sources (e.g.,
Streakwave.com), but I do not sell them or receive any form of compensation
associated with them.
-----Original Message-----
From: village-...@googlegroups.com
[mailto:village-...@googlegroups.com] On Behalf Of Nicolás Echániz
Sent: Thursday, June 26, 2014 3:07 AM
To: village-...@googlegroups.com
You received this message because you are subscribed to the Google Groups
"Village Telco Development Community" group.
To unsubscribe from this group and stop receiving emails from it, send an
email to village-telco-...@googlegroups.com.
To post to this group, send email to village-...@googlegroups.com.
To view this discussion on the web visit
mundi.net.