Hey, this is the first time I have looked into air-stream in about 5 years so I am a bit rusty in regards to what is good, luckily it appears that technology has become a lot better.
I am looking to set up an AP on a 4m mast (highest allowed in marion council on a non residential building without approval). At this point I want the AP to provide free net access to local residents and most likely link into the air-stream network also.
I know that I will want a high gain omni directional antenna for the local residents to tap into and they will most likely use directional antennas.
The AP would gain internet access via another wireless link that will go to my office via a directional antenna on each end.
The AP would hook into air-stream (i assume) using a direction antenna to another node? not sure on this.
What I need to know is, is there a list of recommended antennas?
Also what AP should I use to make this happen? For the 'public' network I would intend shaping the data with some sort of linux box so that they dont flood the link but also so they dont flood the DSL link (I will have a proxy server also), I intend having dozens of users on the system at the same time. At this point I would set up a G network but look to change over to N once it becomes more mainsrteam.
Can anyone recommend an AP that will handle what I want to do? Would I need to go for a managed AP like a Dlink DWL-3200AP (http://www.dlink.com.au/Products.aspx?Sec=1&Sub1=18&Sub2=44&PID=220) or would I be better off using something like a Dlink DWL-G700AP (http://www.dlink.com.au/Products.aspx?Sec=1&Sub1=18&Sub2=42&PID=214) and let the linux box handle everything? I am not hung up on Dlink, just that I have looked at them briefly and can get good prices on them.
Is there any other nodes that are set up similar to the one that I want to set up? If so point me in that direction.
I have read the FAQ's and such but the info tends to pretty general, any specific advice for this setup would be appreciated.
Thanks
Recommendations
Hi there! I recommend against any general "computer brand" wireless access point (Dlink, Linksys, Netcomm etc.) These are all mainly for indoor usage, the minitar days are over. I recommend something designed specifically for outdoors for a number of reasons:
1) Quality and temperature specification
2) Usually better support of open source wireless tools
3) Power specifications - to overcome the noise floor that has risen dramatically
4) Sensitivity specifications
There has been a lot of new products released specifically for outdoor wireless, to the point where some off the shelf all-in-one solutions work out cheaper and less time consuming than the equivalent DIY solution.
Generally you will see Air-Stream people having one of these hardware setups for an AP + backbone
- x86 PC board running two miniPCI cards
- Custom embedded board (e.g Wrap) running two miniPCI cards
- All-in-one AP box + All-in-one backbone box
A simple all-in-one solution is the UltraWAP from freenet-antennas. It comes in a number of different power outputs.
For a custom board or PC board, popular miniPCI cards in order from lowest to highest performance would be CM9, Senao 200mW, Senao 400mW, Ubiquiti SR2.
Also forget 802.11g for anything longer than a couple hundred meters.
My favourite brand for outdoor wireless is Ubiquiti. Their products like Litestation, Powerstation and AP-One are professional all-in-one outdoor solutions. They are not mainstream though so are bit hard to find, for someone starting out I recommend UltraWAPs or Wrap board/PC board with CM9s. I have not personally tested the UltraWAP but there is a lengthy thread somewhere here about them and Rob Clark from freenet-antennas knows everything about them.
As far as antennas go, anyone around here would tell you 8 dBi superpass omni is way to go and for directional the hills 25 dBi or kenbotong 25 dBi (available through bulk buys on occasion) have both been proved reliable.
Other things to take into account is weatherproofing, pigtails, coax cables/connectors, ethernet cable and power supply. These often forgotten items sometimes blow the budget ;)
Similar nodes to you would be Air-Stream-Salisbury-North, Air-Stream-CLG, Air-Stream-Hillbank, Air-Stream-Athelstone.
I am happy to discuss more specific information with you at one of our monthly meetings.
Great Information
Just a note on the Senao 400mW, Ubiquiti SR2 that shadey has recommended these cards are only used on APs with low gain antenna.
Where the CM9 (or equivelent) are used with Dish Antennas for client connections or backbones.
Generally we have also found that Omni antennas with a gain above 8dbi eg 15dbi are not much use for APs as the vertical signal path is too flat and limites the number cleint connections in the meadium range. These 15dbi antenna better suite flat areas and no good for use in very high locations.
Where the 8dbi Omni has consistantly shown to be the best all-rounder for AP applications.
Can't go wrong with a Superpass..
..8 dBi omni. Yep..
Ok, where do I get one? :D I
Ok, where do I get one? :D
I intend running the whole setup indoors until I get the software sorted out.
Definitely helpful Shadey
Definitely helpful Shadey and Robert.
I went to a meeting once at Rose Park, I think at that point there was Skye and about 2 other AP's :D
I will send thru an application for membership today or tomorrow, when the meetings generally occur?
What sort of budget would you guess that I would need to get this up and running, could it be done for under $1k? I deallly I want to spend about $500 but I cant see that happening.
I notice that the general recommendation is to have 2 x minipci cards, with my project wouldn't I need 3? One for the backbone to my office, one for the backbone to air-stream, and one for the local AP? That being the case, wouldn't I need to have a box capable of using 3 cards?
I understand what you said regarding the vertical path of an omni antenna, but I want to have it so that people accessing the local AP can do it using the cheapest possible hardware (not cheap in a cheap sense, cheap in a low cost good value sense), so would I not be better off having a higher power omni and in return the end users could use a lower powered directional antenna?
Hallett Cove is pretty hilly, does that mean that I would want a better vertical spread? moreso, does that mean that I should have a better vertical spread than an omni can offer? and if that is the case, should I be looking at a different antenna all together, say multiple sector antennas?
You mention that G is pretty shite for range, are you therefore saying that I would go for a B solution? or would a B/G solution be better where those in range can use G and those further away can use B? IIRC B is 11mbps, and wireless is a hub technology, so that would mean there would be a lot less bandwidth to play with as opposed to a G solution. I currently get 21mbit over my indoor G setup.
I notice that there is 2.4GHz and 5GHz nowdays, would I be better off with one or the other or both? ie use one for the backbones and one for the local AP.
I was also thinking of powering the setup with solar / battery backup, has anyone else tried this with/without success?
I would prefer it if the entire setup could be controller remotely as where I plan on putting it may not allow regular access.
B vs G and hardware costs
There are a number of ways of doing things..for example your office could just be a client pointing at the omni.
Most of us with solid 802.11b links running at 11 Mbit get real tcp throughput of around 700 KB/s, even as far as 12 km from the omni. I don't know of anyone getting more than that at 2.4 GHz in the outdoor long range by switching to 802.11g. This is for a number of reasons - low level software support, 2.4 GHz noise, better sensitivity in hardware for b, tighter timings in g. As B is more lenient, it would probably perform better in a multiple client situation.
For dedicated point-to-point backbones (the 'core' network) we use exclusively 802.11a links now (running at 5.1 or 5.8 GHz.) For someone wanting to expand the network by providing local coverage, they would connect into an existing 11b AP with a directional antenna which we also call a 'backbone' (e.g. Salisbury-North) even though it is slightly different from a dedicated directional to directional core backbone.
Budget estimates that I like to throw around are $400 for a client, $600 for a network expansion, $900 for 802.11a backbone and AP. This assumes new high quality equipment and factors in all the little bits and pieces. However with the newer all-in-one solutions coming out, a setup could become quite a bit cheaper.
A 'network expansion' is one of the best ways for someone to get involved with the Air-Stream network since it is simply tacked onto existing infrastructure, quickly gets them connected to the network and expands the network coverage.
802.11g Wireless Access Point
The main reason 802.11g does not work well over distance is due to ACK timeout issues.
http://www.air-stream.org/ACK_Timeouts
Unfortunatly ACK timing changes for the 802.11g standard is not supported under any OS. But is in the 802.11a standard under FreeBSD and other linux BSD OS systems, hence why 802.11a is used for backbones. Even if you could change the ACK timeout under 802.11g it would mean it would only then support point to point connection as it is dependant on a specific distance between wireless devices.
Seems straightforward.
Seems straightforward.
Thanks for the reply
Thanks for the reply Shadey.
The reason I want a dedicated directional link to my office is so that I can provide internet access to the AP at full speed (I get 6mbit/1mbit where I am), but I also want to run the proxy server there so I imagine there will be a fair bit of traffic along that link.
I would be happy to just make a network expansion but I assume I wont be able to get net access thru it, hence my reasoning about the office backbone. ALSO I wouldn't want to have my clients running thru the existing backbone and raping it.
$900 seems that it fits under my $1k (soft) ceiling for the project. For the client fitout I would probably go for a bulk order of some sort to get the price down. I would look to have a standard "kit" for the clients, if they cant use the standard kit I would then look at a custom solution. Even @ $400, it is pretty cheap over a couple of years.
Ordering from hills Soon
Hi
I am making an order to hills soon. For a 25dB antenna. If anyone wants to make an order at the same time please let me know.
It wont be for a two or three weeks, i have exams then will be away for a week.