|
1
|
- Drew Dean
- SRI International
- Computer Science Laboratory
|
|
2
|
- I make my living courtesy of IP law.
|
|
3
|
|
|
4
|
|
|
5
|
- What does it mean to “own” something that I’m not allowed to understand
how it works?
- Am I responsible for what my computer does without my knowledge?
|
|
6
|
|
|
7
|
- All DRM systems so far have been designed in secret
- Recent (post-1980) cryptographic solutions have been designed in public
- E.g., the AES competition to replace DES
- The history of closed designs’ security is riddled with failures
- Only the NSA is large enough to do meaningful internal review
|
|
8
|
- It can take nearly forever to find problems in cryptographic protocols
- Needham-Schroeder Public Key
- 3 messages
- 18 years to find the problem!
- While being a standard example in the literature
- DRM faces an even harder problem: Adversary has many more attacks
- Power analysis
- Timing analysis
- Fault injection
|
|
9
|
- New algorithms, modes of operation, protocols, etc. proposed all the
time
- You need serious credentials before you’ll be taken seriously
- Many broken in time for next year’s conference
- Repeat
|
|
10
|
- When things are open, this game works well
- Harder, but possible, in a closed world
- Impossible with DMCA, EU Copyright Directive, etc.
|
|
11
|
- Assertion: We don’t know how to solve the DRM problem today.
- We can’t proceed to play the usual research game
- Hence, we will never solve the DRM problem
|
|
12
|
- Watermarking
- Code obfuscation
|
|
13
|
- Robust Watermarks
- Meant to withstand transformations that leave original recognizable
- Images: scaling, cropping, rotation, etc.
- Sound: transposition, noise, time dilation, etc.
- Fragile Watermarks
- Both: meant to be imperceptible by people
|
|
14
|
- Usage tracking
- Metadata storage
- DRM policy enforcement
|
|
15
|
- Integrity protection of originals
- Detecting lossy compression
- This appears to be solvable
|
|
16
|
- September 2000, 3 weeks
- No documentation
- 4 “robust” watermark technologies
- Devastating results:
- Craver, Wu, Liu, Stubblefield, Swartzlander, Wallach, Dean, Felten,
“Reading Between the Lines: Lessons Learned From the SDMI Challenge,”
USENIX Security Symposium, 2001.
- Stern and Boeuf, “An analysis of one of the SDMI candidates,”
Information Hiding Workshop, 2001
|
|
17
|
- Software is malleable
- Tamper-resistant hardware is rare and expensive
- Can we obfuscate software for better security?
|
|
18
|
- In a completely general way, no
- Barak, et al., On the (Im)Possibility of Obfuscating Programs, CRYPTO
2001
- Cloakware has tried hiding a key in a DES implementation
- Jacob, Boneh, Felten, “Attacking an obfuscated cipher by injecting
faults,” ACM DRM workshop, 2002
- No good, uniform definitions of the problem
|
|
19
|
- Crypto doesn’t just solve the problem
|
|
20
|
- Most security needed for low unit cost, mass market items
- That’s where the big money is
- High unit cost items (e.g. market research reports) have different
business models/needs
|
|
21
|
- Technical problems will persist, but …
- Consumer will pocket veto technologies that fail offer consumers good
value propositions by doing nothing
- An exceedingly simple process for the consumer: keep wallet firmly in
pocket
|
|
22
|
- Critics:
- “Worse, it treats every user like a potential criminal, and tries to
impose new controls on music people paid for years ago. So I actually found it insulting, as
well.”
- “Sony seems so concerned about copyright that it has made getting music
onto the Clip a pain…. Can you imagine Sony product managers sitting
around a conference room, planning to make a product more frustrating
to use?”
|
|
23
|
- Critics:
- “Worse, it treats every user like a potential criminal, and tries to
impose new controls on music people paid for years ago. So I actually found it insulting, as
well.” – Walter Mossberg, The Wall Street Journal, March 2, 2000
- “Sony seems so concerned about copyright that it has made getting music
onto the Clip a pain…. Can you imagine Sony product managers sitting
around a conference room, planning to make a product more frustrating
to use?” – Stewart Alsop, Fortune, February 21, 2000
|
|
24
|
- Technical measures for DRM have a bad track record
- Technical solutions to legal problems are a bad idea
- Legal solutions to technical problems are a bad idea
|