wlan security?

[sinsi]

hutch, as always the voice of reason  :bdg

A question from left-field - does the wireless link run full duplex? or Does that apply to a wireless link?
So much to learn, so little memory...

[sinsi]

OK, old topic but same problem with speed.

Still get slow=wireless, fast=cat5 but I know that.

The weird thing is...installed vmware, installed 2000pro. Went to download ie6sp1 (latest for 2000) from the host (xp home) thinking that it would be quicker
than going through a virtual nic. Got 50-60K/sec, wasn't doing anything else, tried it in the guest (2000) and got 400+K/sec.

Is this some sort of problem with timing things in the guest?

[vanjast]

To drag this topic a leeeeetle further

I've been busy with some homegrown video hardware designs, using DSPs and got sidetracked into encryption (this is a major topic of DSPs).
What I've worked out is that no encryption scheme nowdays is safe(hmmm -- Internet Banking comes to mind  :green2 ), and only the deterrent is the imagination of the encrypter.

What i worked out is a scalable modular DSP pcb, and with the correct scalable firmware, you're well on your way to crack anything...
Maybe what it requires is something simple  :bg

[ecube]

if you're using standard wep+rc4 then it's a cakewake gaining access to your router, you should use a strong protection most routers offer with aes.

[sinsi]

Update - just installed the win7 beta and get 300-400 Kbytes/sec via wireless, must be my XP configuration.

[johncoths]

I guess there are latest wireless network security news, analysis and research from Network World and featured products and services from trusted WLAN Security vendors.. Just searching some site for more information.


_________________
Security Systems

[hutch--]

If you have to route ethernet cable through noisy places, use CAT6, 4 twisted pairs with a hard nylon former to accurately space them and it has a far greater level of noise rejection than earlier CAT5 and 5e.

As far as radio links go, if you can get at both ends of it, roll your own encryption with something fast. On the rare occasion I need to encrypt something I use a random pad at both ends with a crude but fast XOR algo. If you maintain unique pads it cannot be broken and even if you reuse the same pad over time they would have massive problems getting a big enough sample to break it. Its worth using a simple fast encryption before and after a pad to cover up the odd character that is XORRED against a zero or itself.