Obfuscated Circuitry?

ortholattice writes "The current issue of EDN has an article Cunning circuits confound crooks that discusses methods that attempt to foil the viewing of software in embedded designs. Interesting is its view on reverse-engineering, which the article consistently calls theft: "As programmable logic increasingly encroaches on high-volume territory formerly dominated by ASICs, unscrupulous operators are licking their lips at the prospects of easily duplicated, or even reverse-engineered, designs." "...The other harder but possibly even more damaging form of theft is 'reverse-engineering'...""

Comments

Clarification from he who caused all the ruckus;-)

[bdipert]

I understand I'm a bit 'popular' on slashdot.org today ;-) I'll say publicly what I've said privately to everyone who's written me so far; although I admit that I didn't explicitly state this, what I meant was reverse-engineering with intent to illegally re-use IP protected by copyright and/or patent. I even mention legal reverse engineering when I brought up Integrated Circuit Engineering....

It appears from some of the earlier-made postings I've just scanned that some of you figured this out already. I thank you for defending me. And for those who I confused with my less-than-exact wording, I apologize. Profusely. Now quit clogging my email inbox! ;-)

I'll be publishing a print clarification of this point in an upcoming issue of EDN.

Regards,
Brian Dipert

Re:That's right

[buckrogers]

This may sound shocking, but I don't care about the trade secrets of a company where I don't work. I never signed a non disclosure agreement with the company, so I never agreed not to disclose trade secrets.

Patents are designed to give the company a limited monopoly in exchange for them providing the rest of humanity with the information on how that device or procedure works... This is considered _good_ by most because it prevents knowledge from being lost.

However, a company has _no_ such protection for information that they don't share with humanity. Companies who wish to keep their information away from everyone employ a tactic know as trade secrets. This means that they try to keep information secret so that others can't do what they can do.

However this tactic has one serious drawback, other people can learn your secrets through looking at your products, or by simply watching your procedures... Then those other people can compete against you with your own information... We can't have that now!

If something isn't protected by a patent, it is fair game to be reverse engineered and that information used against them in a competetive market... If they feel like sharing then our society will reward them with a 17 year monopoly. Their choice...

Only the whiners go crying to court when their secrets become public knowledge... As if the court can somehow make everyone forget the truth... *laughs* No court is _that_ powerful.

DMCA, reverse engineering, and patents

[Samrobb]

OK... here's an odd question that popped into my head:

The DMCA explicitly allows for reverse engineering for compatability purposes. What if I am interested in reverse engineering a circuit design, piece of code, etc. not for compatibility purposes, but in order to determine if they designer/implementor is infringing on a patent that I or my company holds?

Now, according to the DMCA, I'm a criminal - I've engaged in reverse engineering for other purposes than compatibility. No matter that I may have proof, via the reverse engineering, that someone was infriging on a patent; according the the DMCA, I committed a crime in order to obtain that information.

Am I missing something here? Or does the DMCA - which it's advocates touted as being essential to protecting intellectual property in today's digital world - allow someone to essentially ignore patents under the right circumstances?

well...

[Auckerman]

Some time ago, there was a really big powerful computer company that decided to make Personal Computers and licensed a clone of CP/M from a small company that was owned by the son of a personal friend of the CEO of the larger company. This Personal Computer (PC) had an "protection mechanism" called a BIOS that was designed to stop other companys from making a PC that did the same exact thing as this larger company. Then a smaller company came along and "reverse-engineered" this BIOS, licensed the CLONE of CP/M that the larger company was using and began selling CLONES of the larger companies computers.

They were sued. They won. The WinTel PC industry was born. This doesn't even take into account the amount of software "look and feel" cloning that took place some 8-9 years later.

If figuring out how something works is a crime, then curosity should be outlawed. Quite frankly is someone can reverse engineer it, odds are it was so damn obvious, it didn't deserve protection in the first place.

Right idea, WRONG computer:

[Speare]

The story is right, but you have the wrong computer.

Apple's ROMs had entrypoints that were all over the ROMs' address range, because they didn't want to dedicate any area as a jump table. Franklin copied the ROM verbatim, as rewriting it would screw up the entrypoints. Apple sued Franklin, and WON.

IBM made the BIOS (with function numbers instead of haphazard entry vectors) specifically so that it could be re-written, extended, improved over time. They PUBLISHED the source code to the whole BIOS, and knew that this put them in the risk of being cloned. COMPAQ rewrote the BIOS, function by function, complying with the data interfaces only. IBM sued COMPAQ, and LOST.

I am appalled.

[Joffrey]

As an intellectual property attorney, I am appalled by the stance taken by the author of that article. Below please find the letter I just e-mailed him to briefly correct his misperceptions:

In reference to your article, "Cunning Circuits Confound Crooks," found at: http://www.ednmag.com/ednmag/reg/2000/10122000/21d f2.htm , I find it particularly disturbing that you refer to reverse engineering as "theft." In particular, the following text gives me tremendous pause:

"The other harder but possibly even more damaging form of theft is "reverse-engineering." In that scenario, someone uses the information stored in the programmable-logic device to reconstruct the original circuit details and then alters and incorporates those details in part or whole into other designs." Your analysis is woefully misguided.

Intellectual property protections in the United States exist for copyright, trademark, patent and trade secret information, and each of these areas includes its own particular set of limitations. These limitations are present for very good reasons, ranging from free speech to the encouragement of innovation. For any truly new, useful and unobvious circuit, patent protection can be obtained -- protection that gives the inventor the exclusive right to manufacture, use and sell the patented invention for a limited time. However, that protection comes at a price: the inventor must disclose to the world precisely how to make and use the invention, so that others may build upon it and so that further innovation may be encouraged.

Similarly, trade secret protection also has limitations. Trade secrets are protected only while they are precisely that: secret. Since trade secret protection (a) gives unlimited time-duration protection, and (b) fails to educate the rest of the world and thereby foster further innovation, it is extremely limited. Once a given technology is no longer secret, it may be used by anyone freely.

If a circuit designer decides to forego the greater protections afforded by patent, he or she cannot complain about reverse engineering under the law. So long as someone is not directly infringing a copyright (or mask work) by literally copying a chip design, they are free to use the underlying ideas to improve their own devices.

Reverse engineering is not theft, either legally or ethically, and I suggest that you consider my comments in your journalistic pursuits.

Joffrey X. XXXXXXXX
xxxxx & xxxxx, LLP