Slashdot Mirror
DRM in Real-Time and Embedded Systems
An anonymous reader writes "In this guest column at LinuxDevices.com, Victor Yodaiken speculates on the implications (and potential catastrophic consequences) of Digital Rights Management Passport (DRMP) technology to embedded, real-time, and mission critical computer systems. Quoting from the article: "When a technology gets pervasively embedded in microprocessors, computer boards, and software, it will alter the performance of power turbines, jet engines, medical instruments, cell phones and missile guidance systems. Unfortunately, DRMP technology is incompatible with security and with the kinds of reliability needed in safety critical or mission critical applications.""
DRM in rocket launching chips might indeed have strange effects
... 3 ... 2 ... 1 ... LIFT OFF
Operator:
Launch System: launch operation aborted, you do not have the rights to "the final countdown"
"Son, in a sporting event, it's not whether you win or lose, it's how drunk you get" - Homer J. Simpson
Sir, the missile headed for the terrorist traing camp is changing it's coordinates! It looks like it's targeting the house of a Kazaa user.
I still love the smell of napalm in the morning though.
"Try browsing the Internet without enabling cookies and Java to see how easy it is for pervasive options to become non-optional."
It's a valid point, tho. I like some of the workarounds, such as Opera's willingness to throw out all cookies at the end of the current session, if said options are selected.
Still, the author appears rather alarmist; DRM is a licensing technology, not a security technology, as the author stated. Thus, WHY would consumer-grade "hardware" be found in professional-grade medical hardware? That's like buying a Packard Bell for IBM's web server... it just won't happen.
On that note, it'd be interesting to see if Intel/AMD/MS/blah will try to include DRM in "server" versions of hardware and software...
The only way drm would be included in embedded systems is by law. No manufactor would voluntary put it in for obvious reasons. Wince devices would be another story.
And for Hollywood, Its not like some hacker is going to go into a hospital and turn a resperator into an internet file swapping server and take down the whole media industry. Come on and get real!
Drm will only be in pallidium systems so Microsoft can make more profits by being the gatekeeper of the internet and all multimedia. Infact pallidium is really not a drm sollution in itself according to their faq but will be used to enforce it. Its already in Windows2000 and WindowsXP.
I am sure Fritz will make an exception for many critical embedded systems if he decides to write another insane and unconstitional law. After all the military can not be bothered by drm when their systems monitor nuclear missles. All he cares about is his big fat paycheck by his employers. OOps I meant contributers.
http://saveie6.com/
--it will alter the performance of power turbines, jet engines, medical instruments, cell phones and missile guidance systems.
;)
I can't believe it... a last a positive use for DRM hardware !
Karma cannot be described by words alone.
Thus, WHY would consumer-grade "hardware" be found in professional-grade medical hardware?
Because Fritzie-boy is all hot and bothered to close up the "Analog Hole". That means that NO commodity DSP or processor chips can fail to support DRM. One consequence is that embedded device makers will have to get special exceptions for un-screwed up processors and memory (vastly increasing costs and development time due to red tape). If embedded and real-time manufacturers use commodity parts anyway to control their costs then they'll have to contend with DRM just like anybody else. This is where the defib machine letting someone die on account of a licensing issue comes in.
Remember "professional-grade medical hardware" uses many of the same components as consumer grade hardware. The difference is in how it is configured and even more importantly certified to operate correctly. Mandatory DRM basically means that the well EVERYONE is drinking out of is going to be pissed in by Rosen, Eisner and Fritzie-Boy.
In other news, the US launched a Nuclear Strike against China today.
Hillary Rosen had warned China of the implications of the nation's failure to address music Piracy.
"We warned them there would be severe implications, especially after our merger with the BSA brought software piracy under our jurisdiction."
The RIAA used the Digital Rights override software installed in all US computer systems to launch 12% of the US nuclear arsenal at strategic locations in the piracy prone nation.
"We have to protect the profit margins of the music industry. Musicians have a right to profit from their work, no matter what any one government wants."
When a CNN reporter brought up the potential legal implications of such a move, Ms. Rosen replied, "I don't think that's an issue. If I, or any other member of the RIAA is arrested, the President's pacemaker will automatically disconnect, as will the embedded medical devices in the bodies of half the US Senators. We will simply revoke the digital rights of those devices, thus rendering them inoperable."
"Live Free or Die." Don't like it? Then keep out of the USA
"Just because you *can* do something, it doesn't mean you *should*"
I know I'm an old hippie, but I really believe that if Microsoft and Hollywood spent a fraction of the resources they're throwing at DRM solutions into creating a workable micropayments system for the web, and IP owners started selling their goods at reasonable prices, they'd be minting it in no time.
When VCRs first appeared, Jack Valenti decried them as the spawn of Beelzebub, and foretold the death of the movie industry because of home taping. What happened? They now make more money on VHS and DVD than they do in the cinemas.
And just to prove that piracy *isn't* an issue - the release on DVD of Harry Potter *without macrovision* was the biggest ever DVD release at the time. How come, if everyone was just waiting to pirate it?
To some wild conclusions, the author of that piece linked does.
1. Most military gear does not use off the shelf CPUs. An example - F/A-18E/F - while SuperHornet uses armored Cat-6 cables and PowerPC chips, they are specially made hardened chips for military and commercial sat applications. F/B-22 uses 486s as does F-15E but they are special 486s that come out just for military applications. If you sell a part to the US military for a system, you must produce that system for 15 more years. Since the new F-15Es for the US/Israel/Korea are just delivering now, one can expect 486s without DRM for a while, since F-22 may be in it's current model production until 2011, expect 486s until 2026.
Parts for missiles and PDAs sold to the Military are under the same rules.
2. Medical equipment - Usually use embedded OSes and Dragonball, 486s, ARM or Mot 68000 series chips, not the latest and greatest from Intel/AMD. They sure won't be running Palladium. I found that arguement by the author to be, well stupid.
3 I had another point, but I've got to go to work, and I forgot it. Sorry.
The chips it is embedded in will eventually be cheap enough in price to be used in embedded situations. Even wasting the clock cycles on determining that you're not running a DRM application could be key (probably not with the heart monitor, but with flight control definately).
The secret to all of this is that Intel will most likely have a way to completely and absolutely turn off DRM for the chip, because this really would be unworkable for embedded manufactures. Now the task would be for someone to find that capability and distribute it over the internet. This is highly likely to happen.
Is anyone else out there pissed at the fact that they will actually have to log on to the internet to even use their computer? I mean for broadband its ok, but there are a lot of dialup users out there, who don't need to log in right now to listen to music or to watch DVD's.
A piece of code that runs behind the scenes and can stop the user accessing their data or even stop the machine from working at all. Didn't we used to call those Trojans?
Stephen
"Don't write down to your readers, the only people less intelligent than you can't read" - Sign on Newspaper Office Wall
Not so fast there. With the possible exception of the cell phone, none of the systems you've described have any application whatsoever to digital rights management and the idea that DRM code will "somehow" find its way into every IC / processor, even when such application is utterly useless and contrary to the design constraints (and adds substantial costs) is simply unfettered paranoia. Code doesn't just "appear" by itself and attempts to push meaningless extentions of technology into areas which may risk lives is not going to happen. I can assure you that Boeing's fuel management control systems are not built from parts purchased at pricewatch.com, the differential resonance processor in an MRI isn't a .Net Managed Code resource, and the Navy isn't sourcing on-board trajectory guidance modules from RadioShack catalogs. Legislation that attempts to make that happen isn't going to fly because it would cripple the very industries that rely on technology to succeed and form the heart of Western industry. Even the worst case, the one you've predicted, isn't that bad; we'll just do like we always have -- if they build a higher wall, we build a taller latter. It's simple, really.
Look, I don't want to dismiss your ideas outright. In fact, I share your feelings about DRM -- In its present form it only protects the rights of the corporations, not the rights of the consumer. (In that regard, it should be called "Digital Restrictions Management.") However, this article furthers the same "idea taken to an extreme" paranoia that made people worry whether their car would start Y2K morning.
So relax; take a deep breath and go find something substantial to worry about. There are enough big problems out there without sweating the details of something incredibly unlikely to affect the world in the way you've described.
Remember the 105 year olds getting the letters telling them it was time to sign up for kindergarten? That was a y2k glitch. On 1 January 2000 I visited the US Naval Observatory's Time Site and was informed that the date was 1 January 19100.
A cousin of mine was pulling a low six figure salary from 97 through 99 fixing COBOL systems. The bamks/hospitals/etc spent quite a bit of money fixing the systems.
Best Slashdot Co
Bold is mine. This will not just apply to software, it will apply to everything. Music, books, art, etc. The list goes on. Anything that you create now, even if it is for your own amusement, will be shut down by Digital Restrictions Management. This is just one step in the control of *creation of content*.
Entertainment companies do not want to just control all of their content, they want to control ALL content. You will need to register with 'a third party' for a signature to release your *own works*. Of course, to keep the sigantures from just being owned by 'anyone', they will be prohibitively expensive. You will be unable to compete with the entertainment companies, the software companies, and all others. You won't even be allowed to release your own works of art, music or writing.
Somehow I doubt that a themometer will be allowed to shut down anything, in law or in practice.
It is the independent creation of content that is being threatened, and don't you forget it.
This is why any sort of DRM will ultimately fail for any device that isn't a dedicated media player. In order to be successful, when a DRM device has a fault of some sort, it has to assume that process X isn't authorized to execute. This is the antithesis of mission critical systems, which must never fail.
This is also why Palladium will also fail. Microsoft has said that to be useful, Palladium must run on 100 million machines. In order for it to be useful at all, it must fault towards false negatives (i.e., if it thinks something is wrong, it prevents execution rather than defaulting to execute). Assume that a)Palladium works properly 99.9% of the time and b)that each person tries to run a Palladium enabled program one time per day. Even working 99.9% of the time, there'll still be 100,000 errors per day (and we assumed that each person only tries to use Palladium once in a day, too). Because of the way Palladium works, these errors can't be corrected in house, meaning each person must call Microsoft HomeBase (or internet in, if Palladium lets them) and have the error corrected by a person. This process won't be automated by definition, otherwise it could've simply been part of Palladium itself.
Suppose Palladium shits and dies on you while you're trying to do a presentation of your big proposal? Suppose IIS shuts down your business site on the day after Thanksgiving? This isn't something you can fix yourself, you have to fight 99,999 other people for the phone lines to get the error corrected. There's just too much risk using this sort of scheme even in the business world, much less in mission critical embedded processors.
"I see in the near future a crisis approaching that unnerves me and causes me to tremble for the safety of my country.... corporations have been enthroned and an era of corruption in high places will follow, and the money power of the country will endeavor to prolong its reign by working upon the prejudices of the people until all wealth is aggregated in a few hands and the
Republic is destroyed."
--U.S. President Abraham Lincoln, Nov. 21, 1864
-=[ Who Is John Galt? ]=-
A detachment of special forces is pinned down by enemy fire. The bad guys have found a bug in the special forces target tracking software that allows them to confuse it, maybe by putting out heat sources that are right on the threshold of what is flagged as a target by the software. The good guys fix their program in the field, correct the bug and reinstall. The DRM agent rejects the new software and prints a little message: You have tried to run unlicensed software on this processor.
He underestimates the military, take the Marines for example, they are men who solve problems by eliminating their causes. After the first instance of this happening the word will spread quickly in the software developer community of how a bunch of angry Marines showed up at Microsoft HQ (DMRP division) and rammed armed stick grenades up the developers Rectums before pulling all the pins with a string (Paralell processing).
Only to idiots, are orders laws.
-- Henning von Tresckow
Running virus protection takes processor cycles too, so security == overhead
Yep, and virus-scanners are a big pile of poo too, solving a problem that could better be solved by banning microsoft products.
Ever tried to compile an OS while your virus-checker scans each and every source-file in the entire program, each time you access the file?
DRM would be essentially similar: Although you could get admin on your NT box to turn the virus checker off while you compile, a DRM system would have no such facility (i.e. the administrator would be His Billness) and the system would have that very same requirement of scanning every file you access.
Think about it. Think about how long it will take you to check the certificate of every file in even just the linux kernel. It's some factor-of-ten slowdown or so for a virus checker, and will be similar for DRM.
However, his contention that the only use for Palladium/Trusted Computing Platform technology is DRM is wrong. It could be used, for example, with the Brazilian voting machines, to make sure that what you think is the output from the voting software really is. Without keys protected in hardware, you can't be sure. With TCPA, the output from the software (over the net or on floppy disk) can be signed with a chain of keys right down to the hardware. Without hardware help, there's no way to hide keys on remote systems.
On a less serious note, you could be sure that your opponent in a network game is a person, not a gamebot.
That being said, DRM would still be the #1 use for the technology.
What's wrong with encoding a session identifier in the URL? You don't have to put it in a query string if you don't want to. The entire URL is available for coding state.
Cookies are evil and software architects need to get that through their heads. Unfortunately, many projects are staffed only with developers and application programmers incapable of a deep analysis of anything.