August 26, 2016

Archives for December 2005


Open Thread: SonyBMG Lawsuit Settlement

SonyBMG has reportedly reached a settlement agreement in the class action lawsuits. Alex Eckelberry has posted the motion proposing the settlement, and a brief summary of its terms. The settlement must be approved by a court before taking effect.

We’re on holiday hiatus, so I won’t analyze the settlement now. Feel free to discuss it in the comments, though.

(Update: Discussion had already started in the comments on another post, starting here. It has moved over here now.)

(Update 2: Jim Tyre pointed out that what I had previously called the “settlement” was actually a motion asking the court to approve the settlement. I modified the text to reflect this, and I added a link to the actual settlement.)


Happy Holidays

Barring major unexpected news, we’ll be on hiatus until after the holidays. We appreciate you, our readers, but this is the time of year to spend less time with you and more with our families.

Enjoy the holidays, and we’ll see you in January!


Sony CDs and the Computer Fraud and Abuse Act

We’ve written plenty here about the adventures of SonyBMG, First4Internet, and SunnComm/MediaMax in CD copy protection. Today, I want to consider whether the companies violated the Computer Fraud and Abuse Act (CFAA), which is the primary Federal law banning computer intrusions and malware. A CFAA violator is subject to criminal enforcement and to civil suits filed by victims.

A major caveat is in order: remember that although I have studied this statute, I am not a lawyer. I think I know enough to lay out the issues, but I won’t pretend to give a firm legal opinion on whether the companies have violated the CFAA. Also, bear in mind that the facts are different as to First4Internet (which designed and distributed the XCP software), SunnComm/MediaMax (which designed and distributed the MediaMax software), and SonyBMG (which distributed both software systems but may have known less about how they worked).

There are two relevant provisions in the CFAA. The first one, which I’ll call the “spying provision”, says this:

Whoever … intentionally accesses a computer without authorization or exceeds authorized access, and thereby obtains … information from any protected computer if the conduct involved an interstate or foreign communication … shall be punished …

The second one, which I’ll call the “damage provision”, says this:

Whoever … intentionally accesses a protected computer without authorization, and as a result of such conduct, causes damage … shall be punished …

(“Protected computer” is defined in the CFAA to include nearly every computer at issue here.)

Let’s look first at the spying provision. We know that the programs obtained information from the user’s computer (about how the user used the CD drive) and sent that information across the Net to either SonyBMG or SunnComm. In most cases that would be interstate communication. So the main issue would seem to be whether the companies, in installing their software on a user’s computer, intentionally accessed the computer without authorization or exceeded authorized access.

According to the CFAA,

the term “exceeds authorized access” means to access a computer with authorization and to use such access to obtain or alter information in the computer that the accesser is not entitled so to obtain or alter

In the case of XCP, the software that gathers and sends information only gets installed if the user agrees to an End User License Agreement (EULA), so the company is authorized to access the computer. They might still have exceeded authorized access, if the EULA’s terms did not entitle them to obtain some information that they obtained. Given the vagueness of the EULA language, this seems like a close call. Eric Goldman has argued that a court would give XCP the benefit of the doubt.

Things look worse for MediaMax. The company sometimes installs its software even if the user rejects the EULA. In this case the company is not authorized to put software on the user’s computer or to cause that software to run. But they do it anyway. It’s hard to see how that’s not either accessing without authorization or exceeding authorized access. It looks like MediaMax is in jeopardy on the spying provision.

Sony’s position here is interesting. They shipped the affected software, but they may not have known as much about how it worked. The spying provision applies only if the company accessed the computer (or exceeded authorized access) “intentionally”. If Sony didn’t know that MediaMax installed when the user denied the EULA, then Sony may be in the clear even if MediaMax itself is in violation.

Let’s turn now to the damage provision. This provision covers access without authorization, but doesn’t cover exceeding authorization. As I understand it, this means that you’re not in violation if you had any kind of authorization to access the computer.

The provision also requires that there be “damage”. According to the CFAA, damage includes “any impairment to the integrity or availability of data, a program, a system, or information, that causes loss aggregating at least $5,000 in value during any 1-year period to one or more individuals”. As I understand it, the cost of detecting and mitigating a problem, including the value of time spent by people on detection and mitigation, can be included in the loss. Given that, there can be little doubt that each of these software systems caused damage of more than $5000 total. For example, if a system was installed on 100,000 computers and imposed at least five cents in detection and mitigation costs on each one of those computers, the aggregate damage is more than $5000.

It seems clear, too, that the installation of a rootkit, or the installation of software without permission – not to mention the security vulnerabilities caused by the software – constitutes an impairment to the integrity of users’ systems.

So the main sticking point in the damage provision would seem to be access without authorization. XCP gets a limited authorization to access the computer when the user agrees to the EULA, so they would seem to be okay. But when MediaMax installs despite the user rejecting the EULA, that looks to me like access without authorization. Again, it looks like MediaMax may be in trouble.

The word “intentionally” pops up in again in the damage provision, and again it might protect SonyBMG, if SonyBMG did not know that the software was designed to install without authorization.

There are two more issues regarding the damage provision. The first one is a possible objection from MediaMax, claiming that although the unauthorized installation may have been intentional, the damage was not intentional. As I understand it, courts have rejected this reading of the CFAA, holding that only the access must be intentional, but the statute applies even if the damage was an accident. It’s easy to understand why Congress would have wanted to write the law that way, to say that if you intentionally break in to somebody’s computer, you are responsible for any damage you cause to that computer, even if the damage happens accidentally.

The last issue is whether the companies had authorization to install or run software immediately upon insertion of the CD into the computer, even before the user is presented with a EULA. I think there’s a good argument that the companies ran more software than they were authorized to run in that situation, but it seems like a stretch to argue that they had no authorization to do anything at all. It seems reasonable to allow them to at least run enough software to pop up a EULA. In any case, it would be hard to find $5000 in damage from this behavior.

So here’s my very tentative bottom line: XCP is in the gray area but is probably okay; MediaMax may well be in violation; and Sony’s status depends on how much they knew about what the MediaMax software did. Perhaps a court hearing one of the SonyBMG lawsuits will give us its own analysis.

UPDATE (1:30 PM EST): In the comments, Sam points out an important issue that I missed in writing this post. Even if SonyBMG did not know from the beginning that MediaMax installs and runs without authorization, they did find out about it eventually, and they kept shipping MediaMax discs anyway. So the software’s behavior would seem to be intentional on Sony’s part, at least with respect to those discs sold after Sony learned about the MediaMax behavior. ]


Is DRM Good for You?

Randy Picker, a principled DRM (copy protection) advocate, had an interesting comment on one of my prior posts about the Sony incident. Here’s the core of it:

Assume for now that you are right that DRM leads to spyware; all that means is that we need to figure out whether we should or shouldn’t favor active protection/supervision environments.

That gets us to the central point: namely the fact that consumers don’t want it doesn’t tell us anything about whether it is in the joint interests of consumers and producers. I spent the morning writing my exam and then will have to grade it after the students take it (no grad student graders for law profs). By far and away the worst part of the job, and I certainly don’t want it as part of the job, but that doesn’t mean that it isn’t jointly sensible.

Putting that point slightly differently, consumers may gain more from a DRM world than they would from whatever alternative world emerges without DRM; those subject to restrictions rarely want them but restrictions are frequently welfare maximizing; the fact that one party would like to get rid of the restrictions tells me little (nothing, probably) about whether the restriction is in the joint interest of the parties to the transaction.

It’s true in principle that an arrangement can be unwanted but ultimately good for those on whom it is imposed; but I don’t think that observation matters much in the specific case of CD DRM.

To understand why, let’s look at a case where a similar argument has traditionally worked: copyright. Copyright can be understood as an agreement among all of us that we will not infringe. Even though each of us individually would prefer to use works without paying, we understand that if we all refrain from infringing this increases incentives for authors, leading to the creation of more works we can enjoy. By making and keeping this copyright deal with each other, we come out ahead. (That’s the theory anyway. We all know what happens when the lobbyists show up, but work with me here, okay?)

One of the practical problems with this kind of deal is that each individual can gain by defecting from the deal – in the case of copyright, by infringing at will. If enough people defect, the deal could collapse. This danger is especially acute when it’s technologically easy to defect. Some people argue that this is happening to the copyright deal.

Anyway, what Randy is suggesting is that there might be a similar deal in which we all agree to accept some kind of DRM in order to boost incentives for authors and thereby cause the creation of more works than would otherwise exist. I think that if we weigh the costs and benefits, that would be a bad deal. And I’m especially sure it’s a bad deal for CD DRM. Let me explain why.

First, it turns out to be easy technologically to defect from the CD-DRM deal. Experience with the copyright deal teaches us that when it’s easy to defect, many people will, whether we like it or not.

Second, the costs of the CD-DRM deal seem much clearer than the benefits. Allowing spyware-DRM on our computers will open loopholes in our anti-spyware defenses that will foster more spyware infections. And as we have seen already, spyware-DRM will itself expose us to security risks. That’s the cost side. On the benefit side, we have only the dubious premise that CD-DRM might boost record sales. The costs are more certain, and larger.

The best argument against the CD-DRM deal, though, is that it is inferior to the copyright deal. If we’re going to make and keep a deal of this general type, the copyright deal is the one to pick. Compared to the copyright deal, the CD-DRM deal is a loser: costs are higher, benefits are the same at best, and the deal is just as easy to defect from. If we can’t keep the copyright deal, then we won’t be able to keep the CD-DRM deal either. But more to the point, we shouldn’t make the CD-DRM deal in the first place.

I’ve looked here at the specific case of DRM for CDs, but I think the same argument holds for other types of DRM as well. Leaving aside the mythical side-effect-free DRM systems and perfectly just legal regimes that some DRM advocates dream about, and looking instead at DRM systems and legal rules that could actually exist and how they would work in practice – as I am sure Randy and other principled DRM advocates would want us to do – the available DRM deals look lousy. Certainly they look worse than the original copyright deal.

Now I’m not arguing here that the current copyright deal is perfect or even close to perfect. The copyright deal is under stress and we need to keep thinking about how we might improve it or how we might renegotiate it to work better in the digital world. I’m not certain what the best deal would look like, but I’m pretty sure that it won’t try to lock in any kind of DRM.


G-Men Called on W-Hats for WMVD

[Despite our recent focus on the SonyBMG CD flap, our mandate here at Freedom to Tinker covers infotech and policy generally. So I hope any Sonymaniacs in the audience will forgive me for posting about something else today. (If you need a Sony fix, Bruce Hayden can help.) Regularly scheduled Sony-related programming will resume next week.]

There’s a fascinating story going around about the intersection between virtual worlds and real-life law enforcement. (I have written twice before about this topic.) It started in a virtual world called Second Life, which has 70,000 or so members. There’s a group of in-world characters calling themselves the W-Hats. Stories in the Second Life Herald – a foulmouthed but apparently somewhat trusted virtual newspaper about Second Life – depict the W-Hats as a gang of racist thugs. (The rest of the story I tell here is based on the Herald’s reporting.)

One of the cool things about Second Life is that players can create new kinds of objects, by writing small programs in a special scripting language to describe how the objects should behave, and then launching objects into the world.

Things got really out of hand when the W-Hats created a doomsday device. It looked like a harmless little orb, but it was programmed to make copies of itself, repeatedly. The single object split into two. Then each of those split, and there were four. Then eight, and sixteen, and so on to infinity.

Okay, not exactly to infinity but to billions of copies (after thirty-some generations of splitting), at which point the servers running Second Life crashed, and the whole virtual world was knocked off-line. The W-Hats had created a Weapon of Mass Virtual Destruction (WMVD).

The WMVD was detonated more than once, and on at least one occasion Linden Lab, the company that runs Second Life, contained the damage by taking parts of the world offline as a kind of virtual firebreak.

Last week at an in-world holiday party, Philip Linden, the character played by Linden Lab CEO Philip Rosedale, mentioned that the company had called the FBI about the attacks and had turned over the names of some players. Others at the party reportedly praised the action. But was this justified? Should the FBI get involved in this mess?

It seems to me that they should. A WMVD of this sort is just a fancy denial of service attack, and a deliberate denial of service attack against a large network service looks to me like a crime. It’s possible that the first attack wasn’t meant to crash Second Life – though even if not deliberate it was certainly reckless – but subsequent attacks could only have been intended to cause a crash.

There is some indication that Linden Lab may have banned at least one player temporarily because of the attacks, but there is a limit to the effectiveness of in-world punishment. As James Grimmelmann has argued, the worst punishment available in-world is exile from the world – try to impose a stronger penalty on someone and he will simply exile himself by leaving the virtual world permanently. Real-world punishments can be worse than exile and so stronger deterrence is available in the real world. When stronger deterrence is needed, real-world punishment may be the only option.

Some have argued that players shouldn’t be punished for doing things that the world’s coding allows them to do. But that seems to me to be the wrong rule. For one thing, that’s not how things work in the real world, where you can commit all manner of crimes without violating the laws of physics. And it’s just wrong to think that the virtual world can be coded in a way that allows everything good but prevents everything bad. Any virtual world (if I may be forgiven for that phrase) that is complicated enough to be interesting will probably enable some undesired behavior.

It will sometimes be necessary, then, to appeal to real-world law enforcement to handle bad acts in virtual worlds. In general, there are lots of caveats here – for example, in some worlds, in-world fraud or murder is considered just part of the game; and world-builders shouldn’t run to the FBI over minor problems. But the particular case before us seems like an easy one: the FBI should investigate and, at the very least, use its power to intimidate the perpetrators into behaving better.


Make Your Own Copy-Protected CD with Passive Protection

Here’s a great gift idea just in time for the holidays: Make your friends and relatives their very own copy-protected CDs using the same industrial-grade passive protection technology built into XCP and Macrovision discs.

Passive protection exploits subtle differences between the way computers read CDs and the way ordinary CD players do. By changing the layout of data on the CD, it’s sometimes possible to confuse computers without affecting ordinary players — or so the theory goes. In practice, the distinction between computers and CD players is less precise. Older generations of CD copy protection, which relied entirely on passive protection, proved easy to copy in some computers and impossible to play on some CD players. For these reasons, copy protection vendors now use active protection — special software designed to block copying.

Discs with XCP or Macrovision protection employ active protection in conjunction with a milder form of passive protection. You can create your own CD with exactly the same passive protection by following a straightforward five-step procedure. I’ll describe the procedure here, and then explain why it works.

What you’ll need:

  • A computer running a recent version of Windows (instructions are Windows-specific; perhaps someone will write instructions for MacOS or Linux)
  • Nero, a popular CD burning application
  • CloneCD, an advanced disc duplication utility
  • Two blank recordable CDs

Step 1: Burn a regular audio CD

Start Nero Burning ROM and create a new Audio CD project. [View] Add the audio tracks that you want to include on your copy-protected disc. [View] When you’re ready to record, click the Burn button on the toolbar. In the Burn tab, make sure “Finalize disc” is unchecked. [View] Insert a blank CD and click Burn. Be careful not to infringe any copyrights! For loads of great music that you can copy legally, visit Creative Commons.

Step 2: Add a data session to the CD

Start another Nero compilation, this time selecting the “CD-ROM ISO” project type. In the Multisession tab, make sure “Start Multisession disc” is selected; and in the ISO tab, make sure Data Mode is set to “Mode 2 / XA”. [View] Add any files that you want to be accessible when the CD is used in a computer. You might include “bonus” content, such as album art and lyrics. [View] For a more professional effect, consider adding the installer for your favorite spyware application and creating an Autorun.inf file so it starts automatically. When you’re finished, click the Burn toolbar button. Insert the audio CD you created in Step 1, and click Burn. [View] Nero should warn you that the disc you’ve inserted is not empty; click Yes to add your data files as a second session. [View]

At this point, you’ve created a CD that contains both audio tracks and data files. The data files you put on the CD should be visible in Windows Explorer (in My Computer, right click the CD icon and click Open) and the audio tracks should be rippable with your favorite audio player. To add passive copy protection, you’ll need to modify the layout of the data on the disc so that the audio tracks are more difficult to access.

Step 3: Rip the CD as a CloneCD image file

Make sure the CD you just created is still in the drive and start CloneCD. Click the “Read to Image File” button. Select your drive and click Next. Choose “Multimedia Audio CD” and click Next. [View] Select an easy to find location for the image file and click OK to begin ripping.

Step 4: Modify the image file to add passive protection

The CloneCD image you created in step 3 actually consists of three files with names ending in .CCD, .IMG, and .SUB. The .CCD file describes the layout of the tracks and sessions on the CD. You’ll edit this file to add the passive protection.

Start Windows Notepad and open the .CCD file. Modifying the file by hand would be tedious, so I’ve created an online application to help. Copy the entire contents of the file to the clipboard and paste it into this form, then click Upload. Copy the output from the web page and paste it back into Notepad, replacing the original file contents. [View] Save the file and exit Notepad.

Step 5: Burn the modified image to create a copy-protected CD

Insert a blank CD and start CloneCD again. Click the “Write From Image File” button. Select the image file you modified in step 4 and click next. Select your CD recorder and click Next. Select “Multimedia Audio CD” and click OK to begin burning. [View]

That’s it! You’ve created your very own copy-protected CD.

Now it’s time to test your disc. If everything worked, the files from the data session will be visible from My Computer, but the audio tracks will not appear in Windows Media Player, iTunes, and most other mainstream music players. The CD should play correctly in standalone CD players.

How it works. To see how this form of passive protection works, you can examine the layout of the CD you created. Start Nero and select Disc Info from the Recorder menu. You should see something like this:

(The exact number of tracks you see will depend on how many songs you included.)

Notice that the tracks are grouped into two sessions — essentially two independent CDs burned onto the same disc. Unprotected CDs that combine audio and data files contain audio tracks in the first session and a single data track in the second. The only difference in the passive protected CD you just created is that the second session contains two tracks instead of one.

You added the extra track (shown in yellow) when you edited the disc image in step 4. This simple change makes the audio tracks invisible to most music player applications. It’s not clear why this works, but the most likely explanation is that the behavior is a quirk in the way the Windows CD audio driver handles discs with multiple sessions.

For an added layer of protection, the extraneous track you added to the disc is only 31 frames long. (A frame is 1/75 of a second.) The CD standard requires that tracks be at least 150 frames long. This non-compliant track length will cause errors if you attempt to duplicate the disc with many CD drives and copying applications.

Caveat emptor. Yes, your copy-protected CD is “industrial strength” — XCP and Macrovision employ exactly the same passive protection — but even the pros have their limitations. There are many well-known method for defeating this kind of passive protection, such as:

  • Enhanced software – Advanced CD ripping programs avoid the Windows CD audio driver altogether and communicate directly with the CD drive. Thus, programs such as EAC are able to rip the tracks without any difficulty. – Better CD copying applications, including Nero, support a recording mode called Disc-at-Once/96; this lets them create an exact duplicate of the protected disc even though the last track has an illegal length.
  • Other operating systems – The discs can be ripped with standard software on Macs and on Linux systems. These platforms don’t suffer from the limitation that causes ripping problems on Windows.
  • Magic markers – The famous magic marker trick involves carefully drawing around the outer edge of the CD. This blocks out the second session, allowing the disc to be ripped and copied just like an unprotected CD.

And of course, at any time Microsoft could fix the Windows quirk that is the basis for this technique, rendering it completely ineffective.

Despite these limitations, who wouldn’t enjoy finding a homemade copy-protected CD in their stocking? They’re a great way to spread holiday cheer while preventing anyone else from spreading it further.


Inside the MediaMax Prospectus

Bruce Hayden writes that MediaMax, the company associated with the CD-borne spyware product that Sony has not yet recalled, recently filed a prospectus with the SEC in connection with an upcoming stock offering. In the prospectus, the company is required to describe truthfully its business plans and associated risks. MediaMax’s prospectus is a window into the company’s business practices. It was filed on November 4, about a week before we first reported the security and privacy problems caused by MediaMax.

There’s more interesting material in the prospectus than I can cover here. Bruce Hayden describes some of it. You can read the whole prospectus yourself, but most of it is pretty dry. The most interesting parts are the discussion of business risks (note the conspicuous non-mention of security and privacy risks), and the description of the company’s products. The product description is all I’ll write about here.

Page 30 of the prospectus describes how the MediaMax CD copy protection product works. Remember, this is the company’s own description of its product. Here’s the core of the description:

When the disc is inserted, the auto launch feature will activate the MediaMax program on the second session. Depending on the DRM license implementation, this program is either activated directly or through another program. The program first determines if the LMT Software controls are installed on the computer. If not, or if the disc concerned contains a newer version, it will copy the controls from the disc concerned and will install same. The LMT Software controls consist of two dynamic link libraries. The controls are used by the MediaMax application.

Whenever the second session software is executed, the LMT Software controls will first determine if the content protection device driver is installed on the system. If not, it will extract it from the main LMT Software into a separate file and install it as a standard Windows device driver.

The driver first locates all CDROM devices installed on the computer. Then it polls each device to determine if a new disc has been inserted. If so, it reads various elements of the disc to determine if it is a MediaMax protected disc. It is important to note that the driver is completely idle (without any chance to affect the computer or CD/DVD drives), unless an actual MediaMax disc has been detected. Once detected, the driver will insert itself into the communication stream for that drive to prevent any non-authorized activities. While allowing the computer to access the second session and associated content without any limitations, the driver will interfere when applications try to access the first session only.

When the driver detects that the MediaMax disc is ejected, it will remove itself from the communication stream for that drive and switch back to the polling mode. Several enhancements have been implemented to make it very difficult to locate and/or remove the device drivers.

There are several things to note here. First, in describing the installation process, there is no mention of obtaining user consent, or of the possibility that the user might not consent, or of how the product would cope with a non-consent situation. The description is pretty straightforward: when the disc is inserted, they install the software. So the decision to install without consent seems deliberate.

Second, there is no mention of the phone-home feature, even though websites associated with the product talk about how the feature can be used to display third-party ads.

Third, they brag that “enhancements have been implemented to make it very difficult to locate and/or remove the device drivers.” So the decision to resist uninstallation seems deliberate.

Indeed, they make an even stronger statement elsewhere on page 30:

The software is designed to be completely invisible to users, programs and system components.

This is an exaggeration, but it shows that they do aspire to invisibility. Which is interesting because the only way to be “invisible to users, programs and system components” is to use rootkit methods. So it would appear that MediaMax at least planned to follow First4Internet’s lead in shipping a rootkit.

All of this just confirms what I wrote on Friday about how the technical problems with CD copy protection lead vendors to adopt spyware methods. MediaMax’s description of their own product describes software that installs without consent and resists detection and removal, along with an apparent plan to adopt rootkit methods. MediaMax set off down the road of CD copy protection, and they ended up with spyware.


CD Copy Protection: The Road to Spyware

Advocates of DRM (copy protection) have been keeping their heads down lately, while they try to figure out what went wrong in the SonyBMG DRM spyware fiasco. No doubt they’ll try to explain it away as an anomaly – just a little speed bump on the road to the effective, unobtrusive DRM future that they’re sure will be arriving any day now.

There are some problems with this story. For starters, we’re not talking about a single DRM system – we’re talking about two totally separate systems (XCP and MediaMax), developed by rival companies, both of which turned out to be spyware and to endanger users, in strikingly similar ways. Is this just a coincidence?

Of course it’s not. If we look carefully at CD copy protection as a technical problem, we’ll see why DRM designers are drawn to spyware tactics as their best hope of stopping copying. Let me explain why.

CDs store music files in Compact Disc Digital Audio (CDDA) format, which is easily readable by a wide range of devices. If the music is encrypted or stored in some other tricky format, ordinary audio CD players won’t be able to read it, and the disc will be useless to most customers. So backward compatibility requires that the music be stored in a format that is readable by computer software.

(Technical digression: There are actually small differences between how a computer reads a disc and how ordinary audio CD players read it. So-called passive protection technologies try to exploit these differences by putting things on the disc that try to confuse computers without affecting ordinary players. For our purposes, it will suffice to say that purely passive protection systems are not viable, because computers are not so easily confused. To my knowledge, purely passive CD DRM technologies aren’t being used any more, although some current vendors combine passive protection with active measures. For reasons too boring to go into here, passive protection doesn’t really affect my analysis; and so to streamline the discussion I’ll assume from here on that there is no passive protection.)

If the music is encoded on the disc in a format that any software program can read, the only way to stop programs from reading it is to install software on the user’s computer, and to have that software actively interfere with attempts to read the disc, for example by corrupting the data stream coming from the disc. We call this “active protection”.

For example, suppose the user wants to use iTunes to read the disc. But the DRM vendor wants to stop the user from doing this, because iTunes can be used to make copies of the disc. The active protection software will detect this and will interfere to ensure that iTunes gets a garbled copy of the music.

Here’s the key issue: Active protection only works if the DRM software is running on the user’s computer. But the user doesn’t want the software on his computer. The software provides no value to him at all. Its only effects are to stop him from doing things he wants to do (such as listening to the music with iTunes), and to expose him to possible security attacks if the software is buggy.

So if you’re designing a CD DRM system based on active protection, you face two main technical problems:

  1. You have to get your software installed, even though the user doesn’t want it.
  2. Once your software is installed, you have to keep it from being uninstalled, even though the user wants it gone.

These are the same two technical problems that spyware designers face.

People who face the same technical problems tends to find the same technical solutions. How do you get software installed against the user’s wishes? You mislead the user about what is being installed, or about the consequences of installation. Or you install without getting permission at all. How do you keep software from being uninstalled? You don’t provide an uninstaller. Or you provide an uninstaller that doesn’t really uninstall the whole program. Or you try to cloak the software so the user doesn’t even know it’s there.

Of course, you don’t have to resort to these tactics. But if you don’t, your software will have trouble getting onto users’ computers and staying there. If your whole business model depends on installing unwanted software and preventing its uninstallation, you’ll do what’s necessary to make that model work. You’ll resort to spyware tactics. (Or you’ll quit and go into another business.)

Having set off down the road of CD copy protection, the music industry shouldn’t be surprised to have arrived at spyware. Because that’s where the road leads.


Not Just Another Buggy Program

Was anybody surprised at Tuesday’s announcement that the MediaMax copy protection software on Sony CDs had a serious security flaw? I sure wasn’t. The folks at iSEC Partners were clever to find the flaw, and the details they uncovered were interesting, but it was pretty predictable that a problem like this would turn up.

Security is all about risk management. If you’re careful to avoid unnecessary risks, to manage the risks you must accept, and to have a recovery plan for when things go wrong, you can keep your security under control. If you plunge ahead, heedless of the risks, you’ll be sorry.

If you’re a parent, you’ll surely remember the time your kid left an overfull glass of juice on the corner of a table and, after the inevitable spill, said, “It was an accident. It’s not my fault.” And so the kid had to learn why we don’t set glasses at the very edges of tables, or balance paintbrushes on the top of the easel, or leave roller skates on the stairs. The accident won’t happen every time, or even most of the time, but it will happen eventually.

If you’re a software vendor, your software creates risks for its users, and you have a responsibility to your customers to help them manage those risks. You should help your customers make informed choices about when and how to use your software, and you should design your software to avoid exposing customers to unnecessary risks. Your customers expect this from you, and they’ll hesitate to buy your product if they think you’re leaving the cyberjuice on the corner of the table.

The design of the MediaMax/Sony software is a case study in risk creation. I wrote about these risks two weeks ago:

But even if all [the software’s spyware] problems are fixed, the MediaMax software will still erode security, for reasons stemming from the basic design of the software.

For example, MediaMax requires administrator privileges in order to listen to a CD. You read that right: if you want to listen to a MediaMax CD, you must be logged in with enough privileges to manipulate any part of the system. The best practice is to log in to an ordinary (non-administrator) account, except when you need to do system maintenance. But with MediaMax, you must log in to a privileged account or you can’t listen to your CD. This is unnecessary and dangerous.

Some of the security risk of MediaMax comes from the fact that users are locked into the MediaMax music player application. The player app evades the measures designed to block access to the music; and of course the app can’t play non-MediaMax discs, so the user will have to use multiple music players. Having this extra code on the system, and having to run it, increases security risk. (And don’t tell me that music players don’t have security bugs — we saw two serious security bugs in Sony music software last week.) Worse yet, if a security problem crops up in the MediaMax player app, the user can’t just switch to another player app. More code, plus less choice, equals more security risk.

Sure enough, these risks enable the new attack, which exploits the presence of extra code on the system, and the fact that that code runs with full Administrator privileges.

The biggest risk of all, though, is that the software can install itself without the knowledge or consent of the user. When you decide to install a program on your computer, you take a security risk. But you take that risk knowingly, because you have decided the benefit provided by that program outweighs the risk. If you change your mind about that tradeoff, you can always uninstall the program.

But if you decline the MediaMax licence agreement, and the software secretly installs itself anyway, you will face risks that you didn’t choose. You won’t even know that you’re at risk. All of this, simply because you tried to listen to a compact disc.

Experience teaches that where there is one bug, there are probably others. That’s doubly true where the basic design of the product is risky. I’d be surprised if there aren’t more security bugs lurking in MediaMax.

Sony is still shipping CDs containing this dangerous software.


MediaMax Bug Found; Patch Issued; Patch Suffers from Same Bug

iSEC, EFF, and SonyBMG issued a joint press release yesterday, announcing yet another serious security bug in the SunnComm MediaMax copy protection software that ships on many SonyBMG compact discs. (SonyBMG has recalled CDs that use another copy protection system, XCP, but they have not yet recalled discs containing MediaMax.)

As we’ve written before, the first time you insert a MediaMax-bearing CD into your Windows computer (assuming you have Windows autorun enabled, as most people do), MediaMax installs some software on your computer. Once this initial software is on your computer, you are vulnerable to the new attack. The gist of the problem is that MediaMax installs itself in a directory that anyone is allowed to modify, even users who otherwise run with heavily restricted security permissions. Any program that comes along can modify your MediaMax files, booby-trapping the files by inserting hostile software that will be run automatically the next time you insert a MediaMax-bearing CD into your computer. And because MediaMax is run with full administrator privileges, the hostile program gets to run with full privileges, allowing it to inflict any mischief it likes on your PC.

Alex Halderman has discovered that the problem is worse than the press release indicates:

  • You are vulnerable even if you decline the MediaMax license agreement. Simply inserting a MediaMax-bearing CD into your PC paves the way for an attacker to come along and set a booby-trap. The trap will be sprung the next time you insert such a disc.
  • SonyBMG has released a patch that purports to fix the problem. However, our tests show that the patch is insecure. It turns out that there is a way an adversary can booby-trap the MediaMax files so that hostile software is run automatically when you install and run the MediaMax patch.
  • The previously released MediaMax uninstaller is also insecure in the same way, allowing an adversary to booby-trap files so that hostile software is run automatically when you try to use the uninstaller.

    (These attacks are similar to the exploit described in iSEC’s report, but they involve a different modification to the MediaMax files.)

Because of these problems, we recommend for now that if you have a Windows PC, you (1) do not use the MediaMax patch, (2) do not use the previously released MediaMax uninstaller, and (3) do not insert a MediaMax-bearing CD into your PC.

We have notified SonyBMG and MediaMax about these problems. We assume they will develop a new uninstaller that safely rids users’ computers of the MediaMax software once and for all.

The consequences of this problem are just as bad as those of the XCP rootkit whose discovery by Mark Russinovich started SonyBMG’s woes. This problem, like the rootkit, allows any program on the system to launch a serious security attack that would normally be available only to fully trusted programs.

According to the press release, SonyBMG intends to use MediaMax’s banner ad display feature to warn users about these vulnerabilities. While this is a positive step, it will fail to reach users who have rejected the MediaMax license agreement. This group is at particularly high risk, since they are probably unaware that the software is installed on their computers.

Worst of all, it is impossible to patch the millions of MediaMax-bearing CDs that are already out there. Every disc sitting on somebody’s shelf, or in a record-store bin, is just waiting to install the vulnerable software on the next PC it is inserted into. The only sure way to address this risk is take the discs out of circulation.

The time has come for SonyBMG to recall all MediaMax CDs.

UPDATE (Dec. 9): Sony and MediaMax have issued a new patch. According to our limited testing, this patch does not suffer from the security problem described above. They have also issued a new uninstaller, which we are still testing. We’ll update this entry again when we have more results on the uninstaller.