Slashdot Mirror
New Wireless Security Standard Has Old Problem?
eggboard writes "Wireless security expert Robert Moskowitz, who sits on IEEE and IETF committees on that subject, sent me a short paper on a glaring weakness in the Wi-Fi Protected Access (WPA) protocol that's replacing the weak and broken WEP system well discussed here at Slashdot. His paper, which I've posted here, proves definitively that while WPA itself remains robust and secure, the interface for choosing consumer passwords makes it simple to snarf a tiny bit of network traffic and perform an offline dictionary attack. For Slashdot readers, this probably seems trivial, but because Linksys, Apple, and others are letting users enter My Dog Has Fleas as their passphrase, WPA might be less secure for home users than WEP."
If all it took were a dictionary attack to sniff a password, at least it took that much.
This isn't some simple passthrough that can be gotten through by knowing a couple backdoor passwords, it's a real live algorithm.
But in the end, it's up to the user to enter a password and as long as humans remain humans easy to remember passwords will always be chosen over #HrS2sWmNw/()LggDwMn.
It doesn't matter how easy to break a new system is, it's better than having no security.
I recently took my laptop on a trip across Toronto and in a couple of hours spotted around 60 wireless networks. Around 80% had NO encryption enabled at all. And yes, the most common SSIDs are 'default' and 'linksys'.
So make a system more complex and people won't use it - which defeats the whole object of it.
Jolyon
Please read my Canon EOS tech blog at http://www.everyothershot.com
The important thing here is that this allows for actual security for users smart enough to use good passwords. Even in hex users can enter dumb passwords ("AA AA AA AA AA...").
Hold it, someone correct me if I'm wrong, but doesn't this mean that instead of collecting thousands of weak packets in RFMon you just need to collect one packet from each network and brute force it?
Which method is harder to crack? I'd take WEP. Simply because its takes longer to collect the necessary packets; especially on a smaller network. On a larger network it may work out to be better from a security standpoint for the cracker to start a brute force attack on the packet on a spare computer and let it sit for a few days instead of having him hide a pocket PC with a wifi card in range of the AP for a few days.
"Poorly choosen passwords lead to insecurity."
Well, duh. I didn't need three pages of dense, TLA-obscured claptrap to tell me that.
Speaking as a cryptographer and longtime security geek, this weakness is about as damning as... using a 128 bit cipher that only gives 120 bits of protection. Look at the big picture. Most people don't even use WEP, let alone limit access by MAC address. The average user is SO oblivious to security, sharing passwords, opening .EXE attachments... I'd hate to recall how many times I found things like .rhosts files with '++' in them among career Unix programmers who must have known better. WEP was a semi-broken protocol, TACACS+ was a totally broken protocol, there was no way one could use them without compromising security. Just as nobody can use a number of commercial software products without compromising security.
WPA, on the other hand, is a very well-designed protocol. It is only as weak as its users are careless. And one need not choose "h^Ne#b8SV@,4g%yP" as a password to avoid this attack, any semi-uncommon phrase of 4 or 5 words will do.
I will deal with this problem by threatening users with a nasty note in their personnel file if they choose a sh*t passphrase -- and terminate their wireless access. And yes, I will try cracking the passwords myself, just as I have done with operating system passwords for several years.
I sure wish all my security problems were so simple! At least WPA *can* be secure, unlike the steaming heap of offal that most folks call a desktop operating system.
Okay, so users might pick a password which is less than 20 characters and is dictionary based. Guess what? They always will... Security is a balancing act. If you make security too cumbersome, then users will find a shortcut and abuse it, making it worse than no security. If the spec enforced something like: "passphrases must be at least 128 hex characters" you'd end up with a bunch of passwords which were all "AAAAA..." (or something similiar)
The simple truth is people are lazy. How many passwords do you have? And how many password guarded accounts? I bet even the most diligent of us out there only have a small number of "good" passwords which we use for damn near everything and never rotate.
The problem with WEP was flawed crypto. No matter how good my password was, someone could crack it with unacceptable ease. At least with this new scheme those of us with "good" passwords have a chance.