FCC Rules Open Source Code Is Less Secure

An anonymous reader writes "A new federal rule set to take effect Friday could mean that software radios built on 'open-source elements' may have trouble getting to market. Some US regulators have apparently come to the conclusion that, by nature, open source software is less secure than closed source. 'By effectively siding with what is known in cryptography circles as "security through obscurity," the controversial idea that keeping security methods secret makes them more impenetrable, the FCC has drawn an outcry from the software radio set and raised eyebrows among some security experts. "There is no reason why regulators should discourage open-source approaches that may in the end be more secure, cheaper, more interoperable, easier to standardize, and easier to certify," Bernard Eydt, chairman of the security committee for a global industry association called the SDR (software-defined radio) Forum, said in an e-mail interview this week.'"

Comments

Ain't the gov't great?

[canUbeleiveIT]

Just goes to show how much a bunch of gov't bureaucrats know. Or maybe there just being ass-kissy with business again.

Re:Ain't the gov't great?

[eln]

They believe what the people who give them the most money want them to believe. Welcome to government.

Re:Ain't the gov't great?

[Harmonious Botch]

They are more familar with the idea of secrecy and control than ideas like cooperation and standards.

Amusing

[ebbomega]

Because Security Through Obscurity totally worked for:

MPAA (DeCSS)
Nazis (Enigma)
Xerox (Robin Hood & Friar Tuck)
Microsoft (just about any form of security they've ever had)

and about a billion other examples

Re:Amusing

[Martin Blank]

When the Germans kept Enigma a secret, they did nothing more or less common than anyone else was doing, or still does for the most part. National governments by and large do not leave their communications to AES, but instead use (what they at least perceive to be) more secure methods. NSA still keeps our codes secret, Russia's FSB keeps its codes secret, and the UK's GCHQ keeps its codes secret.

One of the advantages to this is that the limited distribution of a given code can (but does not always) limit the number of attacks against it. Whereas a commercial cipher may result in millions or even billions of ciphertexts to analyze, a government cipher may result in only thousands to work with, and it may be more difficult to determine plaintext aspects of a given document for comparative analysis. It's also generally difficult to get the actual cryptographic hardware without paying someone (either from inside or outside) to steal one.

This doesn't work well at all for the kinds of things that the FCC covers, however. I can generate billions of ciphertexts with known plaintexts for some new wireless system, and I can also do analysis against the electronics involved to look for side-channel attacks. Hiding things for commercial items intended for the general public is fairly pointless.

Side note: I'd not heard of the Robin Hood & Friar Tuck trick. That was some very fun reading. Thanks for brightening my morning a bit. :)

Well, they're technically correct, of course...

[Space cowboy]

If I'm trying to break into some code, and I can read the source code to determine how the author protected it, I'll have an easier job (note: "easier", not "easy") because I can home in on the algorithm the author used. I know whether it's Blowfish, DES, AES, IDEA, or a simple XOR or substitution cipher. I know what pre-encrpytion steps were taken, and what post-encryption algorithms were used.

Let's say that in a moment of insanity, I decided to use a basic XOR encryption routine (create each byte in the encrypted stream by XOR-ing the corresponding source byte with every byte in the password save one, rotating that one as I iterate over the source). This is completely and utterly trivial to crack if you have the source code and *know* the routine I used. It's a repetitive cypher, so it's reasonably obvious unless the password is of significant (a sizeable fraction of the source's length) as well. Note the difference - it's easier with the source code.

Now that's a contrived example - no-one in their right minds would use an XOR cypher, but the same principle applies to harder encryption techniques. If you *know* what system was used to protect the source, you have an advantage over not knowing... Did they gzip the source before encrypting it ? Did they use ZIP, RAR, or 'compress' instead ? Did they XOR to hide the obvious compression header ? Is it inverted (last byte first) or was any other transformation done *before* the encryption stage to try and make it non-obvious that a successful crack had taken place ? These are all "knowns" if you have the source code...

So, yes, it is easier when you have the source code. Security through obscurity is rightly derided, but not because it has no value. It is derided because it leads to the use of insecure encryption methods (small keys, using XOR/whatever instead of proper hard encyption, etc) and the fact that once the obscurity is cleared up, there's no more security. The idea is that if you are sufficiently confident that your encryption is unbreakable, you *can* document how you did it in public. That doesn't mean you *should*.

The point though, and why I disagree with the regulators, is that if you're using hard encryption, it really doesn't matter whether it's *easier*, it's not *easy*. It is in fact still so damn hard, that we're talking "impossible in our lifetime(*)" - the relative comparison makes no sense. It's akin to measuring the height of Mount Everest at 6-month intervals - it's always pretty darn high, though you might find some variance due to snowfall.

So, yes, they're right. But by not considering the (tiny) impact of their conclusion, they have made the wrong ruling.

(*) Modulo the discovery of an easy way to crack the encryption technology, of course.

Simon.

Re:Well, they're technically correct, of course...

[kebes]

You're quite right. Obscurity does provide some level of security, though relying on it alone is a surefire way to have your security cracked eventually. (Whereas things that are cryptographically secure will not be cracked in my lifetime.)

Another way to look at it (especially in the context of open source radio) is that whoever is implementing the security has finite resources (money, man-hours, whatever) at their disposal. For every hour they spend trying to obfuscate the inner workings, that is one less hour spent validating that it is *truly* secure (in the "cryptographically secure" sense). If you instead leverage open-source, then you have code that has been tested and vetted by experts the world over. Suddenly the hours spent on adding obfuscation would be a waste of resources: the code is already so secure that adding the slight additional security of obscurity is a waste of time.

So, while obscurity does provide some kind of security... it is actually the most resource-wasteful form of security (alot of effort for something that eventually gets cracked), whereas the more efficient security model is to focus on things that are fundamentally secure (in which case you may as well use open-source solutions, since you get to take advantage of work already done, and the marginal loss of obscurity doesn't end up mattering).

Why is the FCC regulating security?

[pavon]

I am somewhat perplexed as to why the FCC would need to be regulating the security of consumer devices. For organization that need secure communications, there are already many government and private certifications, that insure this. But why on earth would they restrict consumers from purchasing non-secure software radios if they don't need them?

Is this because they feel that software radios could be hacked to broadcast outside of their certified frequency and power limits? Or because they think they need to protect the public from buying 802.11 routers with crappy WAP implementations?

FCC overstepping their bounds yet again

[Anonymous Coward]

The FCC has absolutely no power to regulate nor any say at all in how software radio or television are implemented.

The FCC commisioners are deluding themselves, again, if they think Congress gave them the power to appoint monopolies.

They have already been slapped down once with regards to the DTV Redistribution Control flag and they're about to be slapped down again.

What's next, washing machines and clock radios?

http://pacer.cadc.uscourts.gov/docs/common/opinion s/200505/04-1037b.pdf

If the Foolish Child Commission can't remember the limits of their power, We the People will be more than happy to remind them, spank them and send them to their 'time-out' corner once again.

Re:Unless, of course, I'm an evil corporation

[Space cowboy]

Oh for [insert deity]'s sake, please don't tell them that... If they actually start thinking through every possible way someone could do harm on a plane, they'll shut down the airlines "for your safety and convenience"...

At the end of the day, the most dangerous thing is an intelligent mind with the goal of doing harm. There is little-to-no way to protect against that, but it's not a politically acceptable truth, so they just make life difficult for everyone and hope for the best [sigh]. The *only* reason for all this is to protect *themselves* from a "you didn't do anything" accusation after the fact.

If people would just accept that life == risk, we'd be a lot better off.

Simon.

Ceteris paribus

[hey!]

"Ceteris paribus" -- assuming "allthings being equal", which they never are.

True, if you have two equally boneheaded pieces of software, then exploits in a the closed one are harder to divine -- not by much, but harder. On the other hand, if you have a piece of software that has survived years of public scrutiny by experts, that is presumptively harder to exploit than something some random engineer ginned up in secret.

Something cannot be widely reviewed (which is the gold standard in security) and secret at the same time. So generally, I think open source represents the best by far and the worst by a little of security possibilities.

The ultimate problem is that broad statements like X is more secure than Y are meaningless. You have to specify the context and threat you are concerned with. Is an open source interpreter burned into a ROM inside of microwave oven more vulnerable than a proprietary interpreter? Well, against what? Same goes for the software radio thing.