Hacker News new | ask | show | jobs
by cromulen 3203 days ago
I'm a bit uneducated when it comes to the cryptography involved, but I'm wondering why people here are so determined DRM can't ever work?

Is it because someone will somehow get a copy and upload to torrents/streaming sites which of course won't have DRM. Thus only potentially annoying legitimate (eg. Netflix) users? Or are there other concerns?

14 comments

In practical terms, DRM obviously can be made to work.

People ideologically opposed to DRM tend to have two blind spots about the DRM service model.

First, they assume that DRM users demand that DRM prevent any copies being made. But that's not true: obviously, any video you show a user in the privacy of their home can be cam-copied. It has even been the case (though it will be less and less the case moving forward) that you could obtain a high-fidelity digital copy. DRM users have always understood that to be the case; what's important is not that copies be impossible, but that they be difficult for ordinary users and, ideally, incur a quality hit. If copies are inconvenient and/or of lower-quality, most of the market will pay for legitimate copies.

Second, and more importantly, DRM opponents assume that the restriction DRM users are seeking is indefinite. But for the most part, content owners are much less concerned about long-term restrictions than they are about the new-release window when their content is most in demand. A DRM scheme only has to survive for a couple weeks to generate immense value for content owners.

From a security and cryptography perspective, a scheme that can be resilient against expert adversaries for a few weeks, or even a framework for minting such schemes on demand, is a commercially reasonable proposition.

People who are not fooled by the ideology of DRM are fully aware that DRM is a legal strategy, not a technology.

That doesn't change the technological harm of DRM. Putting a DRM-shaped hole in web standards makes browsers less secure, less stable, and less maintainable.

iTunes copy protection used to be broken in a few hours, Blu Ray is long since cracked. DRM is neither secure nor cryptographically sound ( http://craphound.com/msftdrm.txt ). The business models that work online keep on being built without DRM.

But DRM remains an irresistible fantasy for corporations who haven't worked out the economics of getting Apple, Amazon or Netflix to add locks to their content.

I don't know a lot of software security people who work on browser security that agree with this. The prevailing sentiment is the opposite: that standardize DRM reduces the attack surface of proprietary DRM down to that of a CDM, rather than full-featured browser plugins. By doing so, EME is improving security, not damaging it.
We could have neither, though.
By what, banning plugins? Now you're asking the anti-DRM people to do exactly what they're angry at the pro-DRM people for doing: preventing people from running a particular kind of program on their computer. It's an incoherent position.
Not banning, simply not providing api for them.
You can have arbitrary exceptions to ensure coherency. There is no reasonable need to have a position without exceptions. Reality isn't that simple.
> DRM is neither secure nor cryptographically sound

That's similar to saying words don't communicate well.

It makes no sense to say that DRM is cryptographically unsound - for the very reason you state, DRM is not a technology.

There have been non-optimal algorithmic choices and weak key-management, but those are entirely separate from saying DRM is cryptographically unsound.

I think the most untenable part of the DRM model is the trust required. All user hardware behind the visual part has to be secure. That is, the manufacturer has to correctly implement any protocols.

This goes (if I'm not mistaken) for the Screen, HDMI cable, GPU and OS. It's a hard balance between keeping consumer happy with their choice and only accepting compliant and capable producers.

The downfall of the DRM is either going to be stripping the legal protections, or competitors without DRM being able to offer a better experience. Sadly, whilst cutting DRM gives a massively better experience, it comes at the cost of control over the customer. I think Netflix really doesn't mind that I can't watch Netflix on Plex. Heck, they might even prefer that.

This is what really scares me about DRM, it enables the building of walled gardens for much more than just media content.

>or even a framework for minting such schemes on demand

This is what video DRM will need. For games, where each game has a somewhat customized version of DRM, taking a few weeks to break is good enough. For video, this will give you a few episodes; but once it is broken it is broken for everything until you roll out a new scheme. Overtime, I would expect the breakers to get more efficient and automated at breaking schemes, while the DRM makers get increasingly lazy and complacent.

This is how Bluray's BD+ scheme work(ed).
Is this why my BluRay player needed to be updated to be able to play newer discs?
One other element: over the longer term, the scheme should be difficult enough to work around such that casual copying is discouraged.
Cryptogtaphically speaking, what is DRM. We want to give an untrusted user access to content, so that they can see it, without giving them access to the content, so they cannot copy it. This is simply impossible in theory. In practice, we can make copying it difficult, possibly resorting to hardware support (so decryption happens in the monitor) and rely on the DMCA to make any attempts to break DRM illegal. Of course non of these solutions actually work in theory; and the farther down the arms race we go, the more annoying it gets to legitimate users.
> Cryptogtaphically speaking,

Honestly, this might be why people aren't on the same wavelength. They're not cryptographically speaking.

Cryptography allows Alice to share a secret with Bob without Carol listening in.

In DRM, Alice and Carol are the same person.

I'm not sure if that's correct or not, but either way, that's in complete agreement with what I just said: they are not speaking in terms of cryptography.
Even if everyone were to replace their hardware for DRM-friendly monitors, you can bet that in a week the market will be flooded by chinese made hardware dongles that allows you to bypass it.
I know content creators. There is a large group that hate their content being copied. That's not going to change.
> I know content creators. There is a large group that hate their content being copied.

And yet, most content creators I know recognize that the "old ways" are dying, and community involvement and other value adds are the way of the future.

Look at youtube (pre ad-pocolypse), twitch, and patreon. DRM isn't where the money is at with content-creation. Spotify, apple-music, et. al. don't pay their artists nearly enough. The future of content is in direct distribution - not billionaire funded recording companies.

Is it content creators or content distributers?

I thought most pushback against easier access to content was from publishers / curators / distributers. They are being made obsolete by digital distribution and are using 'but the poor artists' as an smoke-screen to delay their obsolescence.

DRM doesn't work. Every form of DRM has historically been broken in one way or another. There are a couple of current systems that haven't been broken yet but it doesn't matter; content will be available from unauthorized sources the day they are released for streaming and no technical means is going to stop that.

Even Denuvo, the darling of the major games publishers, is broken on a regular basis. The company acknowledges that their software doesn't last forever (they tout '300 days' on their website as the longest any of their DRM schemes has held up, it usually is broken in a couple of weeks) and that the value is protecting the profits during that period.

DRM does work. Whether Denuvo is successful depends on how it is implemented. Implementations of Denuvo have survived a year+ without being cracked. As you said, DRM often isn't intended to be secure forever, just for long enough that the major sales buzz is over anyway.

You can also think of DRM as branching into online services. DRM commonly required some form of interaction with an online service to validate a key or whatever else. As a result, some DRM is in fact built into the foundation of the software and this is the most successful. Think of World of Warcraft or Everquest. To essentially crack that DRM, you have to recreate the server backend which is a serious undertaking. This has been done, but it took a long time, only applies to older versions and doesn't connect you to the legitimate service.

Steam is an online service and games which use SteamWorks are coupling at least some of their features with the online service, but this is not as complicated as something like World of Warcraft so it is easily cracked. Denuvo attempts to fill that gap by more strongly coupling the software with Steam. As a result, Steam is technically the DRM while Denuvo is helping to enforce it and thus is not labeled as "3rd party DRM" on the Steam Store pages.

There is a balance between the popularity of your software and the difficulty of breaking the DRM. If your software isn't extremely popular, then a crack is not in high demand. If the software is easy to crack, then even if it isn't in high demand it may get drive-by cracked. If it's difficult enough, many people will decide that their time is better spent elsewhere even if it is possible to crack.

As we enter the age of encrypted processing where both the code and data are encrypted during execution on the processor, cracking will be time consuming enough and have enough prerequisites that it simply won't be able to keep up with even the most popular software. It will be a while before this can become mainstream, because unless governments start requiring it, there will always be devices without encrypted processing that companies can't ignore in order to maximize profit. It won't surprise me if we see both governments that require encrypted processing and governments that ban encrypted processing.

Obviously for static media like audio or video this is harder to deal with, but there have been a number of novel solutions to this that just haven't been widely adopted. There are also still a lot of consumers with older hardware that publishers want to target, because not targeting them is worse than the concern of piracy. These are more about mitigation rather than outright making it impossible for a time period.

DRM is broken because sooner or later it will always prevent legitimate access to content.
DRM works! It's very effective at giving control of the playback environment to the copyright cartels.

Unfortunately, a lot of people mistakenly believe the purpose of DRM is to prevent copying. That is the justification, not the goal, which has always been about maintaining control over playback device.

By its very nature DRM of media content makes no sense. How can audio and video be delivered to the user without also allowing the user to save and distribute that content later? At some point, the content must be available in unencrypted form, even if that point is when the content is sent to the display, it can just be captured there.
It could be easy to view but hard to save and distribute. Why can't it? Or do you really think these people have no idea what they're taking about?
Not really. If it's available to view, it's available. Full stop.

There's no way around that. You can't magically change the universe so that content that's viewed can't be captured. It HAS to be converted to analog somewhere. That signal can always be captured and converted back to a digital form that's no longer embedded with DRM.

Best case: you make it marginally more difficult for a mom&pop computer user to copy your content. Anyone with a lick of understanding and 100 dollars to buy some hardware will be able to get it without any problem.

Worst case: you introduce all sorts of unnecessary security holes with poorly written software, that can't be audited (legally speaking), and does absolutely nothing to slow the availability of the content online and for free.

>> It could be easy to view but hard to save and distribute. Why can't it?

> Not really. If it's available to view, it's available. Full stop. Best case: you make it marginally more difficult for a mom&pop computer user to copy your content. Anyone with a lick of understanding and 100 dollars to buy some hardware will be able to get it without any problem.

You literally just agreed with my point though.

The issue you are ignoring is that only one person needs to "save and distribute" it, then the pirated version becomes easier to view than the DRMed version.

The proof in the pudding is that every TV show and film is instantly available online to anyone who can work a BitTorrent client.

No. Netflix is way easier for me than torrenting. I've never pirated music, but I watched a few of torrented movies. However since the ITMS, Netflix and Amazon video become available in my country I never bothered with torrents. I don't even have torrent client on my computer.
Hard to save and distribute only lasts until one party figures out how to get the content in an unencumbered format. Even if it was hard/expensive for them to do so, they can distribute it to the world easily.
I recall reading another comment [1] on this submission about some DRM technology being claimed to protect content for some 300 days in practice. To me that sounds like it's working pretty darn well.

[1] https://news.ycombinator.com/item?id=15279266

Consider Intel ME, which has the full control over your device independently on the operating system. If it somehow works together with the DRM module, the DRM will work.
It kinda does, apparently — but modern movie DRM actually relies on HDCP — sending encrypted video to your monitor and decrypting it there.
The problem is that being "hard to save and distribute" means it doesn't work. It has to be more than hard. Piracy release groups enjoy "hard", and you only have to do the hard part once for a given title (or usually, for a given DRM technology). Once you strip it of its protection, you throw it online and anyone can download it. And now that you've broken that particular DRM technology, your results are repeatable and can be automated, making future releases available even more quickly.

Fundamentally, I can read or write to any byte in my computer's memory. That includes whatever is coming out of the piece of code that reads in the protected file and then passes it to the video decoder or the display.

Until we have computers that can read and write bytes in such a way that not even a user with full privileges on their own machine can access them (keeping in mind, this user can control the kernel and even the hypervisor), this can't be made to work.

The people pushing for DRM, in my experience, really do have no idea what they are talking about. They tend to be industry lawyer types. People who do understand DRM are usually aware that what they are implementing can be broken, but their VP told them to do it anyway so they can make a deal with content distributors.

Also, the DMCA forbids circumventing copyright measures, so they like to have that as a legal tool as well. As long as they tried something, they can go after anyone disabling it, using the legal system.

This is where technologies like TrustZone, Intel ME, Intel SGX, SecureBoot comes - you will have isolated environment, while DRM will be run in another chip or execution level. So you will need either elevate your privilegies via vulnerability, or via hardware reversing techniques. And it is already happening. Computers arent that open anymore, they became walled gardens already.
Yeah, it is creepy that ME/PSP has full control over modern x86 computers. But that's not about DRM — blame enterprise IT management stuff (Intel AMT).

Apparently ME is somehow used on the DRM path but really it's kinda irrelevant — the whole point of modern movie DRM is that the video frames get decrypted on your display.

> Why can't it?

They've been trying to accomplish this for many years now, to no avail. This is a nontrivial problem.

>> It could be easy to view but hard to save and distribute. Why can't it?

> They've been trying to accomplish this for many years now, to no avail.

Any evidence that they haven't had any success whatsoever in making it harder to save/distribute?

It's trivial to get almost any media content you want online for free right now.
I don't think so. See my reply to the sibling comment.
Sure. Search for a recent videogame on pirate bay, download a youtube-to-mp3 browser extension, get the calibre DeDRM plugin. Or, if you really want to understand my point, download and learn how to use a disassembler.

For the casual law-abiding user, it succeeds in making it harder to save/distribute. For everyone else, it's only a matter of time before the DRM is broken.

First and foremost that depends on the content being popular. For example there's a ~30-minute video I paid for on Amazon Videos that I would love to have a copy of so that I don't need an Amazon account and internet access just to play the video, but I can't. Why? Because of their DRM and the fact that I can't find it anywhere else. (Admittedly I haven't directly searched on piracy sites since I'm not into that sort of thing, but Googling didn't pop anything up even on piracy sites, and I would be surprised if it's actually out there somewhere.) Sad reality is it's not even because I wouldn't pay for it; I already paid for it once and I'd even pay for it again if someone sold a copy. But the DRM is "working" in terms of making it too hard to get a copy.

But even if some content is popular, DRM certainly prevents people like me from saving a copy, since I'm not one to go on piracy sites and the only reason I want a copy is so I don't have to get authorization from someone every time I want to play something I already paid for. It might be a dumb business decision since I'd already pay for it before getting a copy anyway, but maybe they think that still makes business sense even if I'd pay for it. Regardless of that it still seems to be working as intended.

Yeah, DOOM (2016) took four months to crack https://torrentfreak.com/denuvo-removed-from-doom-after-game... — and after the crack, the DRM was removed.

And casual users might be law-abiding in the USA, but not everywhere :)

The argument is essentially that in order for you to enjoy the content, the content must be present unencrypted at some point before it hits your eyes and ears. Because of this, it should always be theoretically possible to extract the unencrypted content at that point.
DRM can't work because the math doesn't distinguish between receiving and decrypting content for a legal purpose like viewing it, and doing the same for an illegal purpose like copying and distributing it. If nothing else, there's the analog hole of recording the screen (which could be done in software).

If it did work, that would be a legitimate reason to include these potentially dangerous, un-inspectable DRM features. But it fundamentally can't, so foisting this security risk on everyone seems misguided.

A chain of hardware, all with private keys in a 'secure enclave' / 'TPM' can deliver DRM that is as unbreakable as the private keys are hidden and tampering detection is functioning. Really, the only hole that can't be patched is analogue recording. Realistically, tampering with the screen or cable is probably the easiest target. Though, compromising the TPM would have more widespread effects.
Mainly that. There's no way to show me a video in my own home wherein I can't simply videotape the screen. So even if it's perfectly locked down digitally, it has to be converted to analog at some point for human consumption, and you can always capture that.

But I also disagree that DRM should be baked into "the web" as a standard, especially when most users seem to disagree, the EFF disagrees, and almost half of the committee disagrees.

It's inherently ineffective because for content to be viewable, information about the DRM is widely spread. Consider Blu-ray. HDCP is a huge hassle for anyone trying to do anything with their video content, and it offers next to no security: Every Blu-ray player, TV, and disc producer has to hold the decryption keys, which means it's trivial for the keys to leak out to the public.

And consider the number of hands a piece of media goes through during the various stages of content production. Movies get stolen from anywhere from postproduction studios to the factories that produce the discs.

Adding a layer of DRM on top of the consumer-facing product does almost nothing to prevent the content from reaching pirate sites. And once they're there, the rest of the distribution problem is easy.

> why people here are so determined DRM can't ever work

It is just that you can not have a video for watching it, and not having the video for distributing it. Computers don't work that way, you either have the video or don't.

That said, things are more insidious than what is on the above paragraph. It's computers that don't work that way, so the solution that many people thought about was making computers illegal, and replacing them with closed controlled machines that resemble them in many ways but aren't as powerful.

If you're capable of actually consuming it (by seeing it with your eyes and/or hearing it with your ears), then it is possible to pirate it. Exactly zero DRM technologies today actually prevent piracy, and exactly zero DRM technologies ever will.

Meanwhile, user-hostile DRM can and will (and already does) push users to simply pirate the media in question instead of putting up with the DRM. Users lose, publishers lose, browser vendors lose, and pirates are entirely unaffected.

Publishers win control over playback hardware market which probably outweighs the loss, or they would have stopped doing it.
I think the most simple reasons to tell that DRM can't ever work, are that:

First DRM proponents have gone very far to protect their content. It costs them a lot of money and from a economical point of view it is difficult to see how DRM help them making more money.

Second: The guys who use pirated stuff would anyway not pay if DRM was working, so why the media industry would bother to struggle with them?

To me the more interesting question is why this particular issue is so important that W3C felt compelled to break consensus. I think the answer to that question is a lot more revealing than another "DRM is evil vs you're an unrealistic hippie" fight. Even if we take it for granted that DRM is good... why?
Given that apparently, google (Chrome), Microsoft (IE and edge) and Apple (Safari) supported the proposal, one could defend the decision as necessary to prevent obsolescence. That is, if they did not pass this, they'd just be ignored by 3 out of the 4 most popular browser producers.
My guess is they claim it can't work because they don't want it to work, and figure with enough agreement someone somewhere will be convinced. (Saying this independently of my own views on the matter.)