Windows PC As IEEE 802.11 Access Point

Hi All,

The only limitation would be what 802.11 functions the vendor decided not to support in a device mean to be used as STA-only. [Frankly, if I were designing an 802.11 module, I would simply integrate all functions, and disable as necessary, if even that, but that is another matter.]

I looked around USENET a bit and saw that this question was already asked at least once, and several people responded by asking, "Why would you want to do that?" Well, I still want to do it, and I am certain that it is what I would ultimately want to do, all things considered. I am not a noob - I do research in computer networking, so helpful responses will be helpful. :)

[Must be Windows for now, not Linux].

Thanks!

-Le Chaud Lapin-

Reply to
Le Chaud Lapin
Loading thread data ...

I would say that a Windows PC can't be an access point, but maybe a PC plus an 802.11 NIC can. In that case, you would either only have access to that PC, or it would have to act as a bridge or router. I believe the Internet sharing feature of windows would allow access through to another network connection to the internet.

-- glen

Reply to
glen herrmannsfeldt

Hi Glen,

:|

I am currently reading the 802.11 standard:

formatting link
I was too cheap to pay IEEE for a copy, so I got my information from secondary sources, but now it is freely available. :D It is obvious from the document that those who make WiFi dongles enjoy feast not tasted by many.

The harwdare vendors have the option of including or omitting certain functions that would be provided by an AP. Some vendors took the liberty of simply including every function prescribed in the standard, as it eliminates multiple version syndrome.

Actually, I am not trying to achieve pass-throught. I intend to set up a proprietary DHCP-like system inside a PC, and attach a WiFi dongle to the PC to test it with PDA's, each having their own Wi-Fi transceiver. I simply have to find a full-featured dongle for the PC, and a corresponding driver that exposes 802.11 services primitives to a user-mode API. Microsoft already encourages this scenario using Windows CE, making a PDA act as the AP:

formatting link

-Le Chaud Lapin-

Reply to
Le Chaud Lapin

For all else who are interested, the pseudo-generic name for what I am looking for is "SoftAP".

Google it, and much information will come forth.

-Le Chaud Lapin-

Reply to
Le Chaud Lapin

I appears that the complete package is Zyxel G200 series:

formatting link

-Le Chaud Lapin-

Reply to
Le Chaud Lapin

Search for "Soft AP".

Hint: A clear explanation of what you're trying to accomplish, and what you have to work with, would be helpful.

Reply to
Jeff Liebermann

I'm rewriting TCP/IP. :)

Incidentally, the model that I decided to buy is the Zyxel AG-225H:

formatting link

-Le Chaud Lapin-

Reply to
Le Chaud Lapin

formatting link

If you wish to "use all functions" then the zyxel 200 series are not suitable as they are limited in their encryption capability to WEP 256. As yet I am not aware of any upgrades for WPA/WPA2..., I do have a G-202 and do find this limitation annoying.

Reply to
LR

Sorry, I missed your posting. See the URL's I posted for clues on Windoze Soft AP's.

If you have some spare time, could you also fix the economy, end the war in middle east, end global warming, find a tolerable US presidential candidate, and figure out why my truck is getting lousy gas mileage?

That also only answers half my question. What do you have to work with? If a spare PC, you might consider some Linux firewall/router mutations that are intended to run access points, such as M0n0wall.

If all you want is wireless connectivity, you don't need an AP. You could setup an Ad-Hoc (peer-to-peer) wireless network, where the various wireless clients talk to each other directly. Any out of the box Windoze wireless device will do that.

I'm kinda suprised that you've selected Windoze as your initial platform for wireless global domination. Most such development I've seen is initially done on Linux (because it's better and easier) and later ported to Windoze.

No experience with that model. I'm kinda suprised that you would chose a USB devices for a Soft AP. Soft AP is commonly used on laptops to do man-in-the-middle type attacks at hot spots by creating a fake AP. Also, for various wireless DoS attacks. I would think that you would have selected a PCI or MiniPCI card instead.

Anyway, good luck and please let me know when you can fix my truck.

Reply to
Jeff Liebermann

The AG-225H includes WPA/WPA2.

formatting link
As it turns out, I do not need the data hiding part of security, so if the user chooses, it could be left wide open, since I do end-to-end encryption.

On the matter of allowing strangers to utilize the connection, I let each user decide.

-Le Chaud Lapin-

Reply to
Le Chaud Lapin

I have spare PC's, but I intend to use the same test PC I use for everything else.

Yes, but I wanted an actual AP so that I could create worst-case- scenario according to 802.11 spec, meaning testing all the functions of AP. Will minimize any suprises when I port software to something like WRT54G.

It's a matter of taste. I like the Windows (VMS) threading/process model better, and there is no way I would want to manage a 500 file project using Emacs/GNU/Make. Visual Studio makes it effortless.

The software is completely oblivious to whether a USB dongle is being use or a PCI card is being used. So it makes sense to have soemthing that is easily portable from machine to machine for demonstration purposes.

:)

-Le Chaud Lapin-

Reply to
Le Chaud Lapin

formatting link

You will find that is for normal client operation on not when it is acting as an AP. See para 5.3

Reply to
LR

The bigger they are, the harder they crash. Compile and build on your main PC, but do your testing on something that you don't mind having crash. I don't do any development work, but I've watched it being done. One of my friends keeps an image of his hard disk handy, with which he can overscribble the working drive when his latest experiments go awry.

You won't have any trouble finding things wrong with 802.11. In my never humble opinion, it's a good compromise between the various applications for Wi-Fi, many of which were never even considered possible back in 1997. See WiMax (based on 802.16) for how to do the access point model correctly. Incidentally, I consider the elimination of the hyphen in Wi-Fi a major improvement in WiMax.

Please note the an access point does not know anything about TCP/IP. It runs on layer 2 (MAC) and only uses IP for configuration. Rewriting TCP/IP has no effect on the function of the access point. If you're planning on improving the AP function, you'll need to re-write MAC layer, not the IP layer.

Various ladyfriends have informed me that I have no taste. Therefore, taste does not matter to me, assuming you're not talking about eating the computer.

Open Source Linux wireless and routers have created a large code pool from which you can borrow, steal, plagerize, and license. It's particularly good at the hardware level for things like routers. There's nothing comparable for Windoze.

True if you're re-writing TCP/IP. Not necessarily true if you're rewriting 802.11. While todays wirless chipsets bury most of the physical layer in hardware, there's still plenty of code in the driver to deal with timing, events, error recovery, and performance counters. Some older chipsets did TKIP in hardware, but AES in the driver. Go unto the FCC web site and see what's inside your chosen target. Grab the chipset numbers. Look on the vendors web pile for what's done inside the chip. You get to do everything else in the driver (or wireless manager).

This strikes me as a minor issue and somewhat like the tail wagging the dog. If you're planning to reinvent the wheel, it's best to concentrate on tires, rims, spokes and axles, not on impressing your friends with how well it fits on different vehicles.

No rush. Whenever you're done rewriting TCP/IP.

Reply to
Jeff Liebermann

But one cannot ignore the model prescribed by the access point while thinking about solving major problems with TCP/IP, mobility in particular. I plan to run some test to determine the hand-off time under 802.11 model.

Not sure what you mean by this. I am a minimalist in my engineering, which means I will not create/use something unless I think it is necessary (or rational). For example, if I had started writing my stack on Linux, I would have had to deal with the NDIS wrapper mess. With Windows, I get the real deal.

As far as the access point, it is definitely necessary. Certainly not a minor thing, the mobility problem. And in a room with people who matter, a working demo is far more impressive than hand-waving, and that is what the dongle gives me. The alternative is a hack of WRT54G or something similar.

I am just happy that at least one dongle vendor had the courage to stick to their original plan - make AP out of PC. Makes my job a lot easier.

-Le Chaud Lapin-

Reply to
Le Chaud Lapin

I here people say this, but I have been using my development machine for two decades at least with only one major problem, which is destruction of a registry path do to a bad string function. Every other error cause by my software has always been recoverable. I might use VPC every now and then, but not as a rule. The only exception is device driver development.

Aside from the registry error, there is never a time when software I write can hurt my machine. There are only so many catostrophic paths of execution, and I have no trouble avoiding them.

-Le Chaud Lapin-

Reply to
Le Chaud Lapin

Cabling-Design.com Forums website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.