Slashdot Mirror
Your Own Linux Wireless Access Point
prostoalex writes "Peter Seebach decided to build his own Linux-based 802.11b wireless access point. The article on IBM DeveloperWorks talks about the hardware and software requirements, implementing the operating system on the CompactFlash card, loading Apache and Perl onto the server. The build-it-yourself wireless point is not going to be cheaper than commercially available products, but its educational value is immense." And HaeMaker writes "We have all seen the 802.11b/g booster made from a Pringles can. Well, these guys have taken the idea, put some math behind it to find the optimal can size and have turned it in to a cool product."
Props to Seebs, but exactly HOW is this better than dropping $30 on a Linksys access point?
Cretin - a powerful and flexible CD reencoder
Toshiba Magnia SG20
Certainly, rolling your own is a learning experience but this is hard to beat. Where to you get a switch to put in a roll-your-own box?
Life is the leading cause of death in America.
Did anyone else notice they mentioned the wi-fi sniffers?
:(
little keychain devices that tell you when your in a hotspot.. thats great!
http://store.yahoo.com/directron/etector.html
found a place that has them too since the link on the webpage is already broken
!Sig
I wish I had read this article before going out and purchasing my Linksys BEFW (sorry, can't remember the rest of the model number) wireless access point. It seems to drop connections at random, requiring a reboot of the router by pulling out and reinserting the power cord. The logs it gives are nowhere near as comprehensive as the ones that snort would send me every morning, and while I can configure port forwarding and IPSEC passthrough, etc., I miss the fine-grained control I had over all the options when I had a Linux router.
A mini-ITX case and some assorted hardware and I could have made a much more configurable, sturdy access point for only a little more money. Sigh.
Karma: Excellent Birds (mostly as a result of listening to Laurie Anderson)
Building your own access point with hostap on any given linux box is pretty trivial, and cheap too if you consider you've probably already got the box laying around. All you need is a Prism2 chipset card, and there's a ton of them around.
My company is doing a similar thing, only using mini-PCI cards to keep the box's footprint as small as possible.
It works like a charm, and you can even build security on top of it. hostap supports WEP out of the box (although that's not really security) and MAC address authentication, even via radius!
I've worked extensively with this sort of thing, very useful and if you already have a linux box, much cheaper than an off the shelf access point.
how long before the Heath Company sues about using the name Cantenna , which goes back at least to the 1960s?
The comments have so far been dumb. Let me point out that the real strength in this tutorial is the idea that you can construct a sort of mini-appliance with ease using Linux. Before you downplay this man's efforts, consider that what he is doing is demonstrating one of the key strong points of Linux that will keep Linux around for a long time: Its use in appliances and specialized embedded devices. In fact, this is such a critical area that Microsoft has become more lenient and more "open" (I use the term loosely) with their competing WinCE product.
It's not hard to take some of the author's ideas and twist them around for other custom networking applications, or even make cheap wireless appliances based on the concepts he's presented. Hats off to the man for providing us with a bit of education.
Linksys access points are flaky unreliable pieces of shit.
The original wap11 was OK but the succeeding and current versions are an abomination. And no, the firmware updates don't make a lot of difference.
I eventually binned mine and bought a cheapy 3com one which has been great.
Deleted
Wireless Access Point own You!
Many Thanks,
Luke
The rest of their range is similarly reliable. Bin it and get a 3com.
Deleted
This is throwing an 802.11 card in a computer and configuring it. Start etching some circuit boards, break out the soldering iron and spectrum analyzer, then you can considering it building...
Tired of being "punished" by the Slashdot $rtbl since 2002. I'm now over at http://soylentnews.org/ .
Linux as a platform for network appliance sounds like a marraige made in hell. It's a great desktop operating system, but it doesnt scale well into small places, has a lot of bloat, and tends to consume resources.
Of course, security should always be of great concern, and this is why WinCE has made such headway with internet devices as of late - it was created delibirately to be small, robust, secure and stable - everything youd want from your network.
Of course, it must be noted, that linux was written in C which makes it very good, since C is object oriented.
I don't need no instructions to know how to rock!!!!
A similar project can be found here
:
This post on their forums caught my eye
I have been sucessful in using the ROM image in a CompUSA branded 4 port switch/router for US$39.95. www.compusa.com has computer strores throught the USA. The router, from what I understand, is made by FMI. It uses a Samsung 4510B CPU, ADMtek ADM6996 5 port ethernet switch device, 512kB flash and 4MB SDRAM. The part appears to be able to take up to a 4MB flash. I'm using a 2MB part in a TSOP socket. uClinux boots up to prompt and the ethernet port works.
Using technical knowledge, some neato hardware and Open Source resources I made a Wireless Access Point using Linux.
It was really, really hard but well worth it and the Earth was saved.
The End.
The surprise isn't how often we make bad choices; the surprise is how seldom they defeat us.
The build-it-yourself wireless point is not going to be cheaper than commercially available products, but its educational value is immense.
I guess even users with excellent karma can't be expected to read these days =(
Why is a Super Cantenna more powerful than a home-made Pringles can? linkage
What I want to know is why bother making your own Pringles can.....?
The beauty for those of us who have already made our own WAP is that we can have increased features over commercial WAPs, while using hardware we already have. A simple, cheap, (usually freely obtainable from a relative) 486 is all you need to build a great OpenBSD firewall/webserver/fileserver/WAP.
The cost MAY be slightly more than a commercial WAP. Depends on if you have old NICS lying around. But, you can have the full features and richness of OpenBSDs firewalling features, SSL accounts, VNC if you want it, etc, etc.
Building a custom box just to "learn" is kind of pointless. Put that old 486 hardware to work!
here you go
ya don't fucking say.
>>>>>>>>>>>>>> . In Wi-Fi, a router is usually referred to as a base station, and broadcasts your Internet signal. The router is typically attached to an Internet connection via a USB or Ethernet port. A client card usually refers to a PCMCIA wireless card. The client card is attached to your laptop, or inserted into your PDA's PCI slot. An access point is the radio device that receives the signal from the base station. It attaches to your laptop or computer via USB, PCI slot, or Ethernet. An access point can also mean a physical area (Hot Spot) where you can pick up Wi-Fi signals. >>>>>>>>>>>>>>>>> It is just me, or these definitions are rather shaky? E.g. the PCI slot of a PDA.
If you build a WAP with Linux, Build 5, alter the
code and build a network of wireless access points.
Each access point should have 2 cards, possibly 802.11a for the uplink and 802.11b as the downlink.
This can be accomplished. Say for instance you
build a star shaped network and "home" is in the
center and the host you need to reach is 5 miles
to the east, your routing table would have rules
on each WAP that will relay your IP traffic the the proper destination.
I've thought about this before, and it could be interesting for metropolitan areas:
a mini-distro for AP's that:
- shields the internal (personal) net, from a public, wireless net.
- routes between the private net, the public net, and the internet.
- do proper routing between AP's: it would be nice if there was a way to use your neighbours excess bandwidth to the internet, or use them as a failover when there's an outage just for you.
- provide proper limits: stop providing to the public if you reach 80% of your monthly bandwith limit, prioritize a bit of the bandwith for your personal net (so that other people can't use up all the available bandwith so there's none left for you), maybe only provide your connection to "members" of the network, etc.
- make it available in an easy updatable flash image for homemade and commercial AP's.
- make it configurable for people who have even more interfaces: multihoming with adsl and cable for instance, IR/visible light links/...
The idea would be that while not everyone has the expertise to configure their routers for all these features, everyone who wants to can just reflash their AP with an image, and be able to provide/get wireless services, and be a part of a real emergent wireless network. It'd be interesting for other reasons too: where I live for instance, all upload speed for residential cable/adsl connections is capped at 128kbit. If I want to send files fast to a friend a couple of blocks away, they could be just routed over the wireless network, via a couple of hops: no need to go over the internet - chances are the speed will be a lot better than 128kbit that way. Also by cooperating in a network like this, you'd have increased speed and reliability for all - not everyone uses up their bandwith / monthly volume limit all the time.
Can this also work with DragonFlyBSD?
1. Eat Pringles
2. ????
3. Profit!
The only reason we have the rights we have is that people just like us died to gain those rights. -- Cheerio Boy
Could something like this be used for wi-fi broadband for your home to an ISP? I know that something like this can be done with a flat panel array antenna for over a 10 mile distance.
Correct me if I'm wrong, but using CompactFlash as a storage device on an embedded appliance/application whatever is a bad idea because of the half-life of CF.
I wanted to do something similar a while back, and I wanted small. I figured using a VIA EPIA-M and a CF with CF to IDE I could create a pretty small server. Perfect for Kiosks, but I ended up finding a page all about CF to IDE and why it shouldnt be used as a boot device (yeah like I bookmarked it) and was discouraged.
Anyone have any experience with that?
Im dreaming ofa big bndwdth, That can resist the
My Linux foortprint complete with OS and cramfs
.. and I use it as an interface to the ROBOT.
file system with app and utilities fits under
8MB of Flash! So HA!
WINCE takes up 16MB of RAM and is SSLLLLLOOOOOOWWWWW
And my MK104 embedded controller does not tie up resources while running the Linux kernel. I've got
WiFi, ADC, DAC, Sonar, PWM and Digital Audio controlled by the robot as well as GPS on Serial
and a CMU Vision system on Serial.
And.. Shit.. there's my Zaurus PDA that runs Linux, Wifi , My own custom apps I wrote in QT3
Update your FLASH. My BEFW has been very stable.
I use a laptop, zaurus, and my kids computer upstairs is also conneted to the net by WiFi
Any idea what the range is w/ a Cantenna?
Maybe I should start with "Why not to bother." It's pretty clear that buying an off-the-shelf device is a lot cheaper than building a custom box. The hardware that went into this ran me around $400, and that's not including the possible value of a few hours of my time. Having said that, there are two good reasons to build your own wireless access point.
The first is flexibility and customizability. Want a firewall? No problem. Custom routing? NAT? All easily managed. Custom Web-based configuration? Half an hour's work with CGI scripts. You can add traffic graphing if you like. That off-the-shelf box may have a special Windows-only configuration tool, so it won't be possible to ssh in and change your settings. You won't be able to run your weblog off of the Web server. If there's a problem, you have to wait for a "firmware upgrade" -- which may or may not address your problem.
The other -- and perhaps more compelling -- reason is that it's fun, and it's a good way to learn about the issues you may face running an embedded Linux system. Think of this as a hobbyist's test project, and if you come up with a killer app for it, you might just be able to make a profit anyway.
That's insightful? he didn't even read the front page... Troll extraordinaire.
Yes, but this one goes to 11.
...weekly. That's why I'd rather get a hardware solution for myself.
If it took you this far to realize it was a joke, you are: a genius
It's a great desktop operating system
If it took you this far to realize it was a joke, you are: pretty enlightened
but it doesnt scale well into small places
If it took you this far to realize it was a joke, you are: about average
Of course, security should always be of great concern, and this is why WinCE
If it took you this far to realize it was a joke, you are: fading fast
that linux was written in C which makes it very good, since C is object oriented
If it took you this far to realize it was a joke, you are: breathing my air. Please stop
psxndc
The emacs religion: to be saved, control excess.
Actually most build your own stuff is like this. The prototype or early phase is wicked cheap (Pringles can, for example) so you start jacking around with it ... then you just keep on adding stuff that you need until BAM! $300 access point built from scratch.
... hey, free second machine.
Pretty much the way most of my computers were built. 'Oh neat, I can buy that computer for $300 - a complete working system."
Hmmm - needs a sound card. (cha-ching!)
Hmmm - needs a better video card. (cha-ching!)
Hmmm - needs more memory. (cha-ching!)
Hmmm - needs a bigger hard drive. (cha-ching!)
Hmmm - needs a quieter CPU / case fan. (cha-ching!)
Hmmm - needs faster CPU. (cha-ching!)
Hmmm - needs a CD burner. (cha-ching!)
At the end, only the case, power supply and mobo are original on your uber "$300" machine. And then only maybe.
God help the guy that keeps all the spare parts, then goes out and buys a case, power supply and mobo to install them all into because
Glonoinha the MebiByte Slayer
I certainly have too many old boxes laying around. Yes, building your own access point is much more interesting. And you get to customize so that you have the features that you want. But in the long run it will be a lot more expensive to operate that PC that uses a few hundred of watts of power than an access point powered from a small wall wart that uses a few watts of power. If, like most users, you leave your internet connection up 24/7, you will find that any savings that might exist in building a device from spare parts are grossly offset by the cost of running it.
This is not an agrument for or against building your own, just a warning to those who might read the original post and think they can save money by building their own.
I'm an American. I love this country and the freedoms that we used to have.
Im dreaming ofa big bndwdth, That can resist the
I've tested it myself-
h owto.ht ml
My homemade one made out of a large Pedigree dogfood can works better than the Cantenna.
~50ft more range with a much broader spread.
The building instructions I used are here:
http://www.turnpoint.net/wireless/cantenna
I browse at +5 Flamebait- moderation for all or moderation for none.
Pebble Linux
> I don't want to patch my wireless access point weekly. That's why I'd rather get a hardware solution for myself.
...of course, if the vendor doesn't release a timely patch, then, well...
Software security patch / firmware security patch. What's the difference? If the vendor says you need to patch, then you need to patch, regardless.
I know it's an easy one but I still laughed at it:
"...Unfortunately, ours doesn't come with potato chips..."
The FIC MAAT Server from FIC can serve as a Linux wireless access point.
Specs:
Embedded Linux
Transmeta Crusoe TM5600 500MHz
On-board 128MB SDRAM (512MB max)
2X 2.5' IDE HDD 20GB (120GB max)
4 ports 10/100 Mbps Ethernet Hub Switch
Built-in Wireless Access Point (802.11b)
Dimensions: W246 X D240 X H56mm
Unfortunately, it is rather pricey at $995 MSRP or $645 direct from the FIC California branch. Interesting nonetheless.
My journal has some information on building a WiFi horn antenna from cardboard and aluminum foil. These have much higher gains than the cantennas, but are of course larger.
I think that the exact spacing and element sizing required for high-gain array antennas (such as the yagi) at 2.4 GHz are tougher than many amateur antenna builders can achieve. The horn antenna is easy to make, if you don't mind something larger.
Unless you live in a dorm :-)
I already have a linux router set up, so extending it in to a wireless AP could be a fun project.
and I quote, In order to connect to the VPN, client systems must be running a Windows operating system.
Ironic isn't it? They're using PPTP and RAS to implement the VPN. My guess is, though, you might be able to get FreeSWAN running on it though. It is a 566Mhz Celeron with 128 or 256 of memory. And it appears to already be set up with Telnet. Instead of ssh. On a firewall. More Irony.
You are checking your backups, aren't you?
The thing is, it's easier to write an exploit for an operating system (because it's pretty common and easy to experiment on) that to a hardware solution (closed code, hundreds of _different_ firmwares -- wanna reverse engineer some?).
This is a intresting idea I would buy somethinng like this
Linux is like living in a teepee. No Windows, no Gates, Apache in house.
the heathkit cantenna wasn't really an antenna ... it was a dummy load for testing transmitters. The signal was safely disapated into the oil, not the air. Same name, but different purposes.
Besides, it was never registered as a trademark by Heathkit (ah, the simpler days!), or any other company.
HIV Crosses Species Barrier... into Muppets
Well, these guys have taken the idea, put some math behind it to find the optimal can size...
:)
Which, conveniently, is exactly the same size as a regular Pingles can! (Which explains why they only cost 19.95.
Technically theres no difference between a linux based and windows based access point. Sure you could argue for stability, but the only thing that could bring it down is a system crash (which doesnt really happen in windows unless you're doing something on it other than routing data) or a line failure. He's probably just making it to show Linux's 'supperiority' over windows. When someone claims that they can get 10kb/s over windows with a linux server, they entire Linux community goes crazy.
Note that I am not Linux bashing here, I like and use linux all the time, I'm just pointing out that there is no real reason to make these other than bragging rights.
I love soekris gear. Soren is working on a less full featured unit just for people who want to build a vanilla wAP.
However, if you have a soekris, and you want to DIY a wAP - you should really insert a brain to look at the potential advantages instead of just pulling a linux + hostAP mode whoopdedo.
#1. Flexibility. Since this is standard x86 gear, you get to easily make this an 802.11b AP; but now w/ FreeBSD's ath driver (appropriated for l00n1x by madwifi) you can upgrade it to an a/b/g AP. Leet, suddenly the price is a little more justifiable (considering that Enterasys & Cisco gear that was proportedly going to be upgradeable in such a way is conveniently not now that the newer standards are out).
#2. Try doing something security related; Soekris also offering 3des crypto accellerators that are supported by OpenBSD & more. WEP blows, WPA is a proprietary nightmare; IPSec on the other hand... well that is interesting, and doable - especially with for instance, OpenBSD (FreeSwan for linux, but that's not quite up to the same standards).
#3. Maybe you want to make an access point for a commercial purpose, but don't want to spend an arm and a leg getting a proprietary product to do so, or you find that most of the free offerings kind of blow. Safe yourself grief and frustration with something like wicap (http://www.geekspeed.net/wicap/) Then you still get to potentially draw from the security benefits mentioned in #2 by using OpenBSD; plus you have something that is quick & easy to setup and get working.
Don't just do this to put a prism card into hostAP mode, that is incredibly dull. A waste of time & money. Not to mention, about as useful to write about as another wardriving article.
Newsflash: no one currently selling access points "builds them", according to your statement.
Every single one of them is simply providing custom solutions from stock gear. Just like this article.
you get no sympathy from me for your hardships...
my router is a pentium 133 w/32 MB SDRAM running OpenBSD 3.2 on a 420MB disk
Never had a problem with it. Then again, it's only serving about 10 concurrent LAN connections to the outside at once, so...still, plenty for a home/small office network.
//FIXME: Bad
I've been busting my arse to put together a suitable Linux Wireless Access Point HOWTO for months.
m l
Finally it's in the process of getting published at the Linux Documentation Project.
Slashdot links to some chucklehead's one pager.
http://oob.freeshell.org/nzwireless/LWAP-HOWTO.ht
This guy is right on and that goes double if you're going to be putting the device outdoors.
APs are *stupid* - if you use a unix OS on Soekris you can run current tcpdump + libpcap and *see* low level 802.11b errors. I'm doing this with an eye on a drop in device when troubleshooting is required at an outdoor site.
I am very easy to get along with, but I don't have time to waste being nice to people who are being stupid. -Theo
Check out m0n0wall, its a stripped down FreeBSD router distro, running ipfilter, PHP (CGI version), thttpd, MPD, ISC DHCP server, ez-ipupdate (for DynDNS updates), Dnsmasq(for the caching DNS forwarder), racoon (for IPsec IKE). It has support for prism wireless cards too and has a build specifically for the net45xx. I just set one up at home with a netgear MA311 wireless pci card, makes for a nice router.
You talking about the Heathkit company that went out of business more than a decade ago?
I'm an American. I love this country and the freedoms that we used to have.
This definitely means the end of UPOC, a PROPRIETARY ($$$$) wireless media "platform"- who needs it when you can do it free?
I don't think there is free AP software for Windows
The article mentions, " If you're desperate, you can probably use an old laptop!" Net4521 price is mo than $200. 486 Laptop on Ebay goes for less than $50. With a little work you can use any CF as a HD storage, but you might as well use the disk that comes with the laptop. Combine this with a $40 worth of pcimcia ethernet and wifi cards and you have something that will work for $100 or so. If you use Debian as the OS, you can ssh into it and apt-get upgrade it to keep it far more secure than any dinky off the shelf box. If you want a switch to go along with it, stick a $20 switch on it, but it would be easier to just wifi into it. The article is a great guide to building one of these things, even for desperate people tempted to buy a $60 access point with software you don't know or control.
Friends don't help friends install M$ junk.
The article was your service all along!
Friends don't help friends install M$ junk.
Here, for example? It's embedded Linux, you can buy it in a neat little box, or install it on vanilla hardware, and as well as doing the hotspot stuff you can use several together to build a mesh out of the box.
Virtually serving coffee
Ugh... Why is it you only have mod points when the most interesting subject is something like 'Natalie Portman rulz' :-P
Nice work, Simon...
Especially considering that the card is already FCC approved, so you're halfway there...
I will probably never take the time to do this (out of need) but in no particular order:
That $30 router won't work when a new set of protocols comes out. You'll have to buy a brand new router to support g (if you buy a b one) or if a whole new standard comes out. With this project, you can add a new WLAN card to support the protocol & maintain use of the older ones.
Configurability. I'm sure more technically able people will be able to get their routers to do more useful things than they could with my belkin router. Port forwarding could be improved, as well as DMZ features. (my belkin router doesn't do the former very well and the DMZ feature of the belkin router does not work.) You can also set alerts so that if you find your network being scanned/attacked, the router can alert you immediately, whereas that $30 router doesn't.
This is a neat little project, particularly if you have much of the hardware already. I have a couple of computers I need to set up; if I can turn them into routers (wireless or not) that would make them more useful to me than if they were sitting there collecting dust.
I know who the fuck is this.....
The lunatic is in my head
Thanks for thinking of me
---
It would be nice if they published some specs of the power gain for their commercial cantenna to back up the claims that it is more powerful though.
It looks like similar dimensions to my "Campari" cantenna which I've tried to model the gain for. link.
Comparing its performance to a commercial antenna which I have the spec sheet for suggests the calculations are pretty accurate too.