IEEE Approves 802.11i

Dozix007 writes "IEEE has approved a new wireless security protocol dubbed 802.11i, intended to finally provide sufficient security for wireless connections that users don't need to rely on alternate security layers. The new specification works by using AES encryption in the transceiver itself, encrypting data directly at the level just above the actual radio pulses themselves. That makes it transparent for applications sending data through the radio, so legacy programs running on new 802.11i-compliant hardware will automatically get the benefits of the new protocol without the need for modification."

Ah Finally!

[scosol]

"sufficient" security- hahahahah history teaches us nothing apparently

awesome

[joel2600]

Now try explainging to regular people the difference between a/b/i/g/x and which ones work together, which ones don't and why.

i hope the guys at best buy are up to speed to direct the consumers!

Lack of equipment or how it's supposed to work?

[swb]

IANA wireless expert, but isn't one of the annoying gotchas of 802.11g that the presence of a B client drops all connected nodes down to B speeds?

If I'm remembering that right, then what you're experiencing may not be a lack of standards uptake -- you could be connecting to a ton of 802.11g stations, but somebody's got a B card running.

Is this really a good thing?

[kabocox]

I know some seemless intergrated security is better than having it tacked on afterward. I've always felt that if folks trusted a default security layer to be perfect, they will get burned when the defaul layer is broken. You should always have application encryption of important data. You shouldn't just trust that your pipe will be encrypted. Sometimes those pipes get used by unauthorized third parties that's when having everything else encrypted comes in handy. I'm just afraid folks will switch to the 802.11i and not bother to encrypt any of their data.

Change hardware *again*? No thanks

[jeffmeden]

That makes it transparent for applications sending data through the radio, so legacy programs running on new 802.11i-compliant hardware will automatically get the benefits of the new protocol without the need for modification.

And exactly 0% of the hardware will be backwards compatible. Who trusts data privacy flying across a network anyway? Isnt that what we have VPN, SSH, HTTPS, etc. for? IMHO we have more things to concern ourselves with, like interference countermeasures, signal efficiency, etc. Who is going to switch to a new hardware platform just because it offers a different (read: not necessarily better) encryption method?

Re:Long Time Until it Replaces B/G

[radixvir]

thats probably because for most purposes B is fine. i mean who is going to spend more on G when typical internet speeds never even reach 11Mps? G maybe is fine for the office or home where you are talking to local servers or other clients, but starbucks doesnt need more than a B.

Re:Is this really a good thing?

[DAldredge]

And them not encrypting their data is different for how they do things now?

Re:Actually secure?

[cmowire]

Perhaps.

However, you do have to remember that a lot of classified information that would result in really major problems for many governments travels, encrypted, over the airwaves, on a regular basis. A cryptosystem isn't called secure unless it can't be broken in a reasonable amount of time, even if the bad guy knows your algorythm, and even if the bad guy is able to observe your transmissions.

Basicly, what the entire WEP debacle has shown is that when you are transmitting over the airwaves, the importance of secure encryption increases. And that if you are going to make a widespread standard for encryption, you had better check it out with some folks who know encryption first.

Couldn't this be used by terrorists.

[Anonymous Coward]

Isn't untappable wireless communication at least as much of a threat to homeland security as model rockets [slashdot.org]?

OK, but how does it actually work

[mamba-mamba]

You can't just say oh, it uses AES. AES is a symmetric cipher, which implies that there is a shared session key.

How do the nodes generate and exchange a shared session key? Or do you have to enter an AES key manually before you even hook up? That would certainly lock down the node!

It would be nice if someone posted a link explaining at a medium level how it actually works. I don't want to just go read a draft of the standard, but I wouldn't mind reading a few of the important details.

MM
--

Re:Too many goddamn wireless standards.

[Dun Malg]

Why can't they just settle on one standard and go from there?

That's essentially what's happening already. They settle on a standard, people adopt it. The trouble comes with the "go from there" part. Whenever you "go" anywhere new with a standard, the old stuff is non-compliant, thus requiring a new standard.

Re:Better than IPSec over wi-fi...

[Abcd1234]

How is that a stop-gap? IPSec has one purpose: to protect IP traffic data over an insecure link. Sounds like it fits right into the wifi game. And given that it's a proven standard with many interoperable implementations, it still strikes me as an excellent option for people who wish to secure their wireless transmissions. This is especially true given that 802.11i won't be fully adopted in the market place for at least a year or two.

Besides, there are *many* issues regarding security aside from the wire protocol. As one other posted mentioned, key management is one of these issues. How does 802.11i deal with this? I know IPSec has many different solutions available for key management, meaning I can make it fit into my network infrastructure. How does 802.11i fit into this picture?

Now I'm confused.

[JayJay.br]

Maybe I do not have enough knowledge to know shit about this, but it looks to me that this is a standard for encryption, and it obviously would be public key encryption, and transceivers would exchange public keys to talk.

While this clearly means that now no one can sniff the SSID, is this going to be any better for those who leave it at the default? And without any kind of MAC authentication or network protection at upper levels, would knowing the SSID the only difficult imposed against abuse of the network?

Not trolling, I just want to know if stupid admins can still mess this one up.

Re:Key Management

[DdJ]

I'll be really shocked if it works in a way fundamentally different, from a user experience standpoint, than today's systems.

This means I'd bet someone $20 that it'll use a single shared key across the entire network, and client machines will obtain it from a user-entered password.

But since it uses AES, all sorts of people will get excited and believe it's secure.

So I see this as little more than a marketing ploy.

Is it more secure than WEP and WPA? Yes. Yes, it's more secure, because in order to get the password that lets you get on the network and steal network resources and intercept everyone's data, you'll need to run a key logger or watch over someone's shoulder or get a virus on to their machine instead of just watching network traffic.

Re:Sure but does it require new equipment

[tmasssey]

Three things:

1) It's not likely that the 200MHz CPU in that thing is going to handle 54Mbit worth of traffic. AES is not the easiest to calculate...

2) Even so, it's highly likely that a firmware update could *possibly* add this. Will Cisco? My guess is no: they are not incented to make your current device more useful. They'd rather sell a new device.

3) The beauty of OpenSource is that you can add whatever features you want... [seattlewireless.net]

Re:Sure but does it require new equipment

[paranode]

Don't see any reason why a device would need to be created specifically for 802.11i.

Ah, that would be because corporations are greedy. Sure they could give you a firmware upgrade, but they could also peddle a completely new product that costs you money.

Re:Key Management

[DdJ]

No, I'm saying that it's a marketing ploy be cause it is in actual fact not better than the solutions out there today, but decisions were made and messages were communicated merely to create the impresssion that it is. But I was a bit subtle about saying that, which can sometimes be a mistake in public forums.

If you rely on encryption that behaves like that, you're foolish and will have problems.

If you believe this is better than what has come before, you are more likely to rely on it.

Therefore, I actually think this will in practice cause more harm than good with regard to actual security.

IMHO, we need totally wide-open unencrypted wireless, with IPSec and nothing else running on top of that, with secure apps running on top of that. I think any crypto at this layer is essentially smoke and mirrors.

Re:Sure but does it require new equipment

[tmasssey]

According to this article [commsdesign.com], the speed of encryping 128 bits of data with a 128-bit AES key is 730 cycles on a 32-bit MIPS processor. To keep it consistent with your numbers, that's actually >45 cycles/byte. At approximately 5 Million bytes/sec (54Mbit wireless), and 45 cycles/byte, that's 225 Million cycles per second right there. IIRC, the processor that's embedded in the router has a single pipeline at 200MHz, or, at best, 200 MIPS.

In other words, assuming *zero* processing overhead, we're 25 MIPS short for wire-speed encryption.

These are very rough numbers, but think of it this way: do you think Cisco (or whoever) spec'ed a processor substantially faster than what they needed? From my peronal experience, embedded processors do not usually have more than a few percent more performance than they need: rarely do they have even 30% more performance than they need. Even if they design a system with a way-fast processor, one of two things happen: their code bloats to use that speed (or they quit optimizing because they don't need to), or they end up buying a lower-cost, slower processor for production!

In short, it's highly unlikely that the Wrt54g will have anywhere near the CPU power to do wire(less)-speed AES at 54Mbit. Half that? Maybe, but not all of it.

Re:The way things ought to be

[Vengeful weenie]

Just run IPSec over your network. Fixed.