Is cracking WPA/TKIP faster than WPA2/AES?

[hashweasel]

Hi.

I was wondering whether brute force cracking of TKIP is faster than cracking AES. I have a few networks here that use WPA/TKIP and I'm wondering whether it makes any sense to switch them to WPA2/AES.
Is it worth it? Will it make cracking more difficult?
I'm asking about oclHashcat specifically, since it seems to be the only one that can use GPU to crack WPA and WPA2.

Thanks.

[Mangix]

I'm probably wrong on this point as I have never seen a performance difference, but cracking WPA1 should be faster.

As far as switching is concerned, you definitely should as WPA2-AES is required to get 802.11n speeds on most hardware. AES also uses less resources on most hardware since it's easy to implement AES in hardware(ASIC).

edit: just tested, same speed.

[hashweasel]

Thanks a lot!

[atom]

Wait, oclHashcat-plus is not cracking the cipherstream, it's cracking the handshake. Therefore, it makes no difference in terms of speed if it's WPA1 or WPA2.

[hashweasel]

Wait, oclHashcat-plus is not cracking the cipherstream, it's cracking the handshake. Therefore, it makes no difference in terms of speed if it's WPA1 or WPA2.

So what does it do exactly? I mean, in terms of cryptography, what kind of computation is this exactly? How exactly is that handshake encrypted/hashed?

[Mangix]

http://en.wikipedia.org/wiki/IEEE_802.11..._Handshake

Basically, PBKDF2(HMAC−SHA1, passphrase, ssid, 4096, 256) with another 3 or 4 iterations for the last step.

[atom]

The slow PBKDF2 part is same for both, which is why they have the same speed. After that, WPA1 uses MD5 while WPA2 uses SHA1 but thats so fast compared to the slow part that you wont see a difference.