Netgear USB (WG111v2) -- WEP key registry value

I'm doing something a bit out of the ordinary -- programmatically setting
up wireless adapters on Win98 machines. The technique works fine for
Belkins, but Netgear has a dramatically more complex registry environment
for their adapter.

It looks as if my principal problem is that the key is not stored in
plaintext. Normally that's a plus, obviously. Not in this case. I have no
idea how to store a value that will make the adapter work.

I've regmoned the heck out of this thing while making he changes using the
systray widget. I've compared before/after exports. And still, no settings
I make to the registry -- no matter how much they appear to mimic what the
widget does -- enable device operation.

I believe if I could generate the string that's stored here:

HKLM\System\CurrentControlSet\Services\Class\Net\0 002\DefaultKey0

.. . . i'd be a happy camper.

Has anyone else done the kind of bit-level tinkering I'm trying to do,
here, and met with success? This really kind of sucks. The presumption by
manufacturers that everyone on the planet will want to click on a systray
widget to configure their devices, is a frustrating thing for a sysadmin to
observe on the part of OEMs.

--

Scott

> here, and met with success? This really kind of sucks. The presumption by
> manufacturers that everyone on the planet will want to click on a systray
> widget to configure their devices, is a frustrating thing for a sysadmin to
> observe on the part of OEMs.

Sorry but no. If you're a sysadmin you should stop using a protocol
that was broken ages ago and move to something that doesn't require
deploying a symmetric key!

Take a look at any RADIUS based solution using dynamic keying, it's not
new stuff.

David.

David Taylor opined thusly on Aug 11:
>> here, and met with success? This really kind of sucks. The presumption by
>> manufacturers that everyone on the planet will want to click on a systray
>> widget to configure their devices, is a frustrating thing for a sysadmin to
>> observe on the part of OEMs.
>
> Sorry but no. If you're a sysadmin you should stop using a protocol
> that was broken ages ago and move to something that doesn't require
> deploying a symmetric key!
>
> Take a look at any RADIUS based solution using dynamic keying, it's not
> new stuff.

We're using that for our domain machines, and it's darned nice (pushing out
certs with policy). The legacy usage is for homebrewed winterms using a
custom shell atop Win98SE. I'd welcome tips for implementing Radius on
Win98 where we don't want to authenticate users. :-/

--

Scott

> custom shell atop Win98SE. I'd welcome tips for implementing Radius on
> Win98 where we don't want to authenticate users. :-/

How much money have you got cos I can give you AES256, FIPS140-2
certified (end to end), non spoofable device authentication with no user
auth from anything from DOS all the way up.

If you don't want to do RADIUS then why not just go for WPA? WEP is
long dead for the purposes of encryption.

David.

David Taylor opined thusly on Aug 11:
>> custom shell atop Win98SE. I'd welcome tips for implementing Radius on
>> Win98 where we don't want to authenticate users. :-/
>
> How much money have you got cos I can give you AES256, FIPS140-2
> certified (end to end), non spoofable device authentication with no user
> auth from anything from DOS all the way up.
>
> If you don't want to do RADIUS then why not just go for WPA? WEP is
> long dead for the purposes of encryption.

Understood. Again, if I can code for that without a lot of pain for an
alternative shell running on Win98, with drivers for an adapter that's
cheap, I'd be happy. Programmatic config of the machine has to be pushed.

--

Scott