High Speed Digital Backbone Radio Link
page has received
since May 1 1999
This page outlines my experimental aims towards
building a low cost, high speed digital radio point to point
link. Such links are intended as a backbone for a future high
speed data network, and will run at speeds above 1 Mbps. Intended
frequency of operation is in the amateur 10 GHz and/or 24 GHz
Why a high
speed data network?
I see several reasons why a high speed data
network can be of benefit to the amateur community. Besides the
obvious reasons of experimentation and relieving the congestion
on the current packet radio system, I have outlined my thoughts
on the matter in my essay: Amateur Radio - A
of High Speed Link:
To fulfill its role and meet the needs of the
amateur community, the digital link should ideally meet the
- Operate at speeds > 1Mbps (faster =
better). Sufficient bandwidth to support multiple
simultaneous voice, data and other connections.
- Range > 20km, preferably > 50km
(line of sight), to allow links between mountain tops.
- Low cost and readily available components.
- Ability to link to end user access points
(hubs, nodes, etc) and existing analogue and digital
- Utilise standard protocols to enable rapid
deployment and development.
- Implement Quality of Service in the
routing algorithms to enable optimal voice/video
performance, even under heavy traffic (kernel hacking
anyone? :-) ).
To achieve these goals, the following
technologies are being investigated:
- Linux operating system. Linux is a freely available
operating system. Because Linux and the majority of
applications which run under it are released with their
source code, this is an ideal development platform. Linux
is also extremely stable and can be customised for
specific applications (like a mini router for use at a
- Ethernet. Ethernet is a standard, well
known networking system, normally used in LANs. While
probably not ideal for radio based systems, surplus
Ethernet cards are easy to come by and driver support is
generally very good. Speeds other than the standard 10
Mbps are being considered.
- Gunn diode transceivers. While not the
latest and greatest, Gunn diode transceivers are
relatively simple for newcomers to setup and offer
reasonable modulation bandwidth, so are a good candidate
for the first generation link transceivers while more
specific hardware is being developed. With a reasonably
sized dish antenna, they are capable of achieving ranges
in the 10's of km region.
- TCP/IP and multicasting. TCP/IP is already
an established networking protocol suite and numerous
applications exist. In addition, TCP/IP supports
multicasting, a technology which may prove useful for
linking multiple repeaters and creating 'virtual
repeaters' on the digital network.
The digital link will be developed in several
stages, ideally by many amateurs working independently, but
towards a common goal. I envisage many of these stages occurring
in parallel, as they are not necessarily serially dependent. Some
of the development to be done includes:
- Publicity and presentation of project to
interested amateurs, Linux hackers and other potential
interest groups. (it would be nice to gain a few new
amateurs along the way :-) ).
- Establishment of basic specifications and
- Sourcing of various components.
- Installation and configuration of Linux on
- Modification and testing of Ethernet cards
at the initial design data rate of 2 Mbps.
- Construction and testing of suitable
microwave radio transceivers.
- Construction and testing of a complete
link on the bench.
- Testing the link under field conditions.
- Configuration and optimisation of
- Development and testing of specialised
software components ('virtual repeaters', linking
- Promotion of system to the amateur
community for full scale implementation.
- Further development to increase speed and
provide new services.
- Last, but definitely not least, to provide
some means of medium to high speed user access via radio.
This is a separate project not addressed directly in this
A lot of the above does not require
transmitting on air, so non amateurs have plenty of scope to
participate, and will be encouraged to obtain their licence so
they can enjoy the fruits of their labour. Oh, and amateurs will
be encouraged to get comfortable with Linux. It is a 2 way street
after all, and everyone has expertise to share. :-)
July 4th 1999:
- Recent developments have changed the
project slightly. A new transceiver design has been
developed by John, KE5FX, which
offers the full 10 Mbps Ethernet bit rate on 10 GHz. As a
result, cards will not need to be modified, which removes
one possible source of uncertainty. Still awaiting the circuit
details at this stage. This project may be one of the
first tests of the new design over long haul (> 5 km)
- A parallel line of research has started
due to several non amateurs playing around with wireless
network devices and high gain antennas (within the limits
of the class licence for these devices). Ranges of
several kilometres are being reported with these cards by
several operators. Depending on the specific devices,
some of these cards may be useful for lower speed links,
as they utilise the 13cm band and offer a neat package.
Some older models are quite cheap (and supported under
Linux), but the hobbyist demand has dried up the supply a
bit. On the downside, there is little, if any
interoperability between cards from different
manufacturers (unlike Ethernet), so the pairs used for
each link would need to be matched.
June 7th 1999:
- Contact has been made with other microwave
operators, and I am following leads for sourcing suitable
parts to build the transceivers. Consideration is being
given to building the first prototype on 23cm, possibly
through modifying FMTV transceiver designs.
- Focus is mostly on generating interest in
the project and seeking others with similar aims to pool
resources. One amateur has indicated interest in working
on a medium speed system which may be suitable for end
user access to the network. Several others have indicated
interest, as well as some non amateurs who have Linux
- The VK-SHF mailing list
has been created as a forum for Australian microwave
operators. Discussion of the high speed link will be
May 27, 1999:
- A minor setback. One of the network cards
I was testing with went up in a cloud of smoke. Took 3
days to remove the stench from the shack! Fortunately,
there are plenty of spares available.
May 8th 1999 and earlier:
- Project has been publicised via the web
(you're reading this now), on the Melbourne Linux Users
Group mailing list, via word of mouth on various voice
repeaters around Melbourne, newsgroups and IRC. Response
has been generally positive with some offers of help,
especially from the Linux community.
- I have drawn up some very basic specs,
mainly from cost/convenience criteria. These are subject
to change as the project takes shape.
- Linux is already running here, so the test
platform is ready. Initial hardware testing will be done
on 386, 486 and Pentium machines running Red Hat Linux
4.2 and 5.2.
- Surplus Ethernet cards have been sourced,
as well as 4 MHz crystals for modification to 2Mbps
operation. Testing of the modified NICs is expected in
- Ethernet cards have been modified to run
at 2Mbps. Hard wired testing is in progress.
Currently, the following areas need more
investigation in the near term:
- Sourcing of suitable microwave hardware.
- Involvement of volunteers and delegation
- Further publicity and promotion.
- Ideas and expertise to develop end user
access systems to the high speed network. Low level link
layer design and programming is way out my area of
Got some ideas or know some helpful links?
If you're interested, would like to help or
simply offer feedback, feel free to email me at firstname.lastname@example.org, or discuss the project and other microwave related
issues on VK-SHF.
Home Page - An excellent site for
Australian microwave information.
Amateur Radio Page
Higher Speed Packet Page
European Microwave News
KE5FX's Page. Includes a design for 10
Mbps link on 10 GHz.
page, using microwave links based on
N6GN's original design.