Published on

  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide
  • `
  • RFID Adhesive Labels
  • Asymmetry result of passive tag vs. battery tag.

    1. 1. 6.857: RFID Security and Privacy November 2 nd , 2004 Massachusetts Institute of Technology Computer Science and Artificial Intelligence Laboratory
    2. 2. Talk Abstract and Outline <ul><li>Abstract: What is RFID, how does it affect security and privacy, and what can we do about it? </li></ul><ul><li>Outline </li></ul><ul><ul><li>RFID Introduction, History, and Applications </li></ul></ul><ul><ul><li>Security Threats and Adversarial Model </li></ul></ul><ul><ul><li>Countermeasures </li></ul></ul>
    3. 3. What is RFID? <ul><li>R adio F requency Id entification: Identify physical objects through a radio interface. </li></ul><ul><li>Many different technologies called “RFID”. </li></ul><ul><li>Others types of auto-ID systems include: </li></ul><ul><ul><li>Optical barcodes </li></ul></ul><ul><ul><li>Radiological tracers </li></ul></ul><ul><ul><li>Chemical taggants </li></ul></ul>
    4. 4. RFID System Primer <ul><li>Three Main Components: </li></ul><ul><li>Tags, or transponders, affixed to objects and carry identifying data. </li></ul><ul><li>Readers , or transceivers , read or write tag data and interface with back-end databases. </li></ul><ul><li>Back-end databases correlate data stored on tags with physical objects. </li></ul>
    5. 5. RFID Adhesive Labels 4 cm
    6. 6. An RFID “Smart Shelf” Reader
    7. 7. System Interface Reader 01.203D2A.916E8B.8719BAE03C Tag Database Reader Network Data Processing
    8. 8. RFID History <ul><li>Earliest Patent: John Logie Baird (1926) </li></ul><ul><li>“ Identify Friend or Foe” (IFF) systems developed by the British RAF to identify friendly aircraft. </li></ul><ul><li>Both sides secretly tracked their enemy’s IFF. </li></ul><ul><li>How do you identify yourself only to your friends? </li></ul>Don’t shoot! We’re British! Oh. We’re British too!
    9. 9. Digression #1: Related Military Applications <ul><li>IFF still used today for aircraft and missiles. Obviously classified. </li></ul><ul><li>Could envision an IFF system for soldiers. </li></ul><ul><li>Lots of military interest in pervasive networks of cheap, RFID-like sensors. </li></ul><ul><li>Monitoring pipelines, detecting biological agents, tracking munitions, etc. </li></ul>
    10. 10. Commercial Applications <ul><li>Early Applications: </li></ul><ul><ul><li>Tracking boxcars and shipping containers. </li></ul></ul><ul><ul><li>Cows: RFID ear tags. </li></ul></ul><ul><ul><li>Bulky, rugged, and expensive devices. </li></ul></ul><ul><li>The RFID Killer Application? </li></ul>
    11. 11. Supply-Chain Management (Not Gum) <ul><li>First Universal Product Code scanned was on a pack of Juicy Fruit gum in 1976. </li></ul><ul><li>Every day, over five billion barcodes are scanned around the world. </li></ul><ul><li>But barcodes are slow, need line of sight, physical alignment, and take up packaging “real estate”. </li></ul><ul><li>Over one billion RFID tags on the market. </li></ul><ul><li>Example: Gillette’s “shrinkage” problem. </li></ul>
    12. 12. Modern RFID Applications <ul><li>Supply-Chain Management </li></ul><ul><ul><li>Inventory Control </li></ul></ul><ul><ul><li>Logistics </li></ul></ul><ul><ul><li>Retail Check-Out </li></ul></ul><ul><li>Access Control: MIT Proximity Cards. </li></ul><ul><li>Payment Systems: Mobil SpeedPass. </li></ul><ul><li>Medical Records: Pet tracking chips. </li></ul>
    13. 13. Prada's RFID Closet MIT Prox Card
    14. 15. Tag Power Source <ul><li>Passive: </li></ul><ul><ul><li>All power comes from a reader’s interrogation signal. </li></ul></ul><ul><ul><li>Tag’s are inactive unless a reader activates them. </li></ul></ul><ul><ul><li>Passive powering is the cheapest, but shortest range. </li></ul></ul><ul><li>Semi-Passive: </li></ul><ul><ul><li>Tags have an on-board power source (battery). </li></ul></ul><ul><ul><li>Cannot initiate communications, but can be sensors. </li></ul></ul><ul><ul><li>Longer read range, more cost for battery. </li></ul></ul><ul><li>Active: </li></ul><ul><ul><li>On-board power and can initiate communications. </li></ul></ul>
    15. 16. Functionality Classes Ad Hoc Networking Active Read/Write Smart Dust 4 Environmental Sensors Semi-Passive Read/Write Sensor Tags 3 Data Logging Passive Read/Write Electronic Product Code 2 Identification Only Passive Read-Only Electronic Product Code 1 Article Surveillance Passive None Anti-Shoplift Tags 0 Features Power Source Memory Nickname Class
    16. 17. Operating Frequencies 3 meters 10-20 centimeters 10-20 centimeters Typical Range 10 meters 3 meters 3 meters Maximum Range? 868-956 MHz 13.56 MHz 120-140 MHz Frequency Range UHF HF LF Range Class
    17. 18. Asymmetric Channels Reader Tag Eavesdropper Forward Channel Range (~100m) Backward Channel Range (~5m)
    18. 19. Security Risks: Espionage <ul><li>Corporate Espionage: </li></ul><ul><ul><li>Identify Valuable Items to Steal </li></ul></ul><ul><ul><li>Monitor Changes in Inventory </li></ul></ul><ul><li>Personal Privacy </li></ul><ul><ul><li>Leaking of personal information (prescriptions, brand of underwear, etc.). </li></ul></ul><ul><ul><li>Location privacy: Tracking the physical location of individuals by their RFID tags. </li></ul></ul>
    19. 20. Espionage Case Study <ul><li>The US Food and Drug Administration (FDA) recently recommended tagging prescription drugs with RFID “pedigrees”. </li></ul><ul><li>Problems: </li></ul><ul><ul><li>“ I’m Oxycontin. Steal me.” </li></ul></ul><ul><ul><li>“ Bob’s Viagra sales are really up this month.” </li></ul></ul><ul><ul><li>“ Hi. I’m Alice’s anti-fungal cream.” </li></ul></ul>
    20. 21. Security Risks: Forgery <ul><li>RFID casino chips, Mobil SpeedPass, EZ-Pass, FasTrak, prox cards, €500 banknotes, designer clothing. </li></ul><ul><li>Skimming: Read your tag, make my own. </li></ul><ul><li>Swapping: Replace real tags with decoys. </li></ul><ul><li>Producing a basic RFID device is simple. </li></ul><ul><li>A hobbyist could probably spoof most RFID devices in a weekend for under $50. </li></ul>
    21. 22. Security Risks: Forgery <ul><li>Mandel, Roach, and Winstein @ MIT </li></ul><ul><li>Took a “couple weeks” and $30 to figure out how produce a proximity card emulator. </li></ul><ul><li>Can produce fake cards for a few dollars. </li></ul><ul><li>Can copy arbitrary data, including TechCash. </li></ul><ul><li>Could read cards from several feet . </li></ul><ul><li>(My card won’t open the door past a few inches.) </li></ul><ul><li>Broke Indala's FlexSecur “data encryption”. </li></ul><ul><li>(Just addition and bit shuffling. Doh.) </li></ul>
    22. 24. Security Risks: Sabotage <ul><li>If we can’t eavesdrop or forge valid tags, can simply attack the RFID infrastructure. </li></ul><ul><li>Wiping out inventory data. </li></ul><ul><li>Vandalization. </li></ul><ul><li>Interrupting supply chains. </li></ul><ul><li>Seeding fake tags – difficult to remove. </li></ul>
    23. 25. Adversarial Model <ul><li>Can classify adversaries by their access. </li></ul><ul><li>Three levels of read or write access: </li></ul><ul><ul><li>Physical: Direct access to physical bits. </li></ul></ul><ul><ul><li>Logical: Send or receive coherent messages. </li></ul></ul><ul><ul><li>Signal: Detect traffic or broadcast noise. </li></ul></ul><ul><li>Can further break down into Forward-only or Backward-only access. </li></ul>
    24. 26. Adversarial Model: Attacks <ul><li>Long-Range Passive Eavesdropper: </li></ul><ul><ul><li>Forward-Only Logical Read Access. </li></ul></ul><ul><ul><li>No Write Access. </li></ul></ul><ul><li>Tag Manufacture/Cloning: </li></ul><ul><ul><li>No Read Access/Physical Read Access. </li></ul></ul><ul><ul><li>Physical Write Access. </li></ul></ul><ul><li>Traffic Analysis: Signal Read Access. </li></ul><ul><li>Jamming: Signal Write Access. </li></ul>
    25. 27. Adversarial Model: Countermeasures <ul><li>Countermeasures will degrade an adversary’s access. For example: </li></ul><ul><li>Encryption degrades logical read access to signal read access. </li></ul><ul><li>Authentication degrades logical write to signal write access. </li></ul><ul><li>Tamper resistance can degrade physical read to logical read access. </li></ul>
    26. 28. Is it really that bad? <ul><li>Maybe Not. </li></ul><ul><li>Tags can only be read from a few meters.* </li></ul><ul><li>Will mostly be used in closed systems like warehouses or shipping terminals. </li></ul><ul><li>Can already track many consumer purchases through credit cards. </li></ul><ul><li>Difficult to read some tags near liquids or metals. </li></ul><ul><li>Can already track people by cell phones, wireless MAC addresses, CCTV cameras, etc. </li></ul>
    27. 29. But…the customer is always right. <ul><li>The public perception of a security risk, whether valid or not, could limit adoption and success. </li></ul><ul><li>Similar to Pentium III’s unique ID numbers. </li></ul><ul><li>Successful boycott of Benetton. </li></ul><ul><li>Privacy advocates have latched on: </li></ul><ul><ul><li>“… e-mails sent to the RFID Journal…hint at some of the concerns. ‘I'll grow a beard and f--k Gillette,’ wrote one reader”, Economist Magazine, June 2003. </li></ul></ul><ul><ul><li>“ Auto-ID: The worst thing that ever happened to consumer privacy”, CASPIAN website. </li></ul></ul>
    28. 30. Digression #2: RFID Public Relations <ul><li>The industry never misses a chance to shoot itself in the foot. </li></ul><ul><li>“ Track anything, anywhere”. </li></ul><ul><li>“ Wal-Mart Caught Conducting Secret Human Trials Using Alien Technology!” </li></ul><ul><li>Lesson: If you don’t want people to negatively spin your technology, don’t make their jobs easier. </li></ul>
    29. 31. Security Challenge <ul><li>Resources, resources, resources. </li></ul><ul><li>EPC tags ~ 5 cents. 1000 gates ~ 1 cent. </li></ul><ul><li>Main security challenges come from resource constraints. </li></ul><ul><li>Gate count, memory, storage, power, time, bandwidth, performance, die space, and physical size are all tightly constrained. </li></ul><ul><li>Pervasiveness also makes security hard. </li></ul>
    30. 32. Example Tag Specification Anti-Collision Support Random Number Generator Features 10 μWatts Power Consumption per Read Passively powered via RF signal. Tag Power Source 10,000 clock cycles. Cycles per Read 100 read operations per second. Read Performance 3 meters. Backward Range 100 meters. Forward Range UHF 868-956 MHz. Operating Frequency 200-2000 gate equivalents. Security Gate Budget 1000-10000 gates equivalents. Gate Count 32-128 bits of volatile read-write memory. Memory 128-512 bits of read-only storage. Storage
    31. 33. Resource Constraints <ul><li>With these constraints, modular math based public-key algorithms like RSA or ElGamal are much too expensive. </li></ul><ul><li>Alternative public-key cryptosystems like ECC, NTRU, or XTR are too expensive. </li></ul><ul><li>Symmetric encryption is also too costly. We can’t fit DES, AES, or SHA-1 in 2000 gates. </li></ul><ul><li>(Recent progress made with AES.) </li></ul>
    32. 34. Hash Locks <ul><li>Rivest, Weis, Sarma, Engels (2003). </li></ul><ul><li>Access control mechanism: </li></ul><ul><ul><li>Authenticates readers to tags. </li></ul></ul><ul><li>“Only” requires OW hash function on tag. </li></ul><ul><li>Lock tags with a one-way hash output. </li></ul><ul><li>Unlock tags with the hash pre-image. </li></ul><ul><li>Old idea, new application. </li></ul>
    33. 35. Hash Lock Access Control Reader Tag metaID ← hash(key) metaID Store ( key,metaID) metaID Who are you? Store metaID Locking a tag Querying a locked tag Unlocking a tag key metaID = hash(key)? “ Hi, my name is..”
    34. 36. Hash Lock Analysis <ul><li>+ Cheap to implement on tags: </li></ul><ul><li>A hash function and storage for metaID . </li></ul><ul><li>+ Security based on hardness of hash. </li></ul><ul><li>+ Hash output has nice random properties. </li></ul><ul><li>+ Low key look-up overhead. </li></ul><ul><li>- Tags respond predictably; allows tracking. </li></ul><ul><li>Motivates randomization. </li></ul>
    35. 37. Randomized Hash Lock Reader Tag: ID k Knows tag ID 1 ,…, ID n R,hash(R, ID k ) Query? Select random R Unlocking a tag ID k Search hash(R, ID i )
    36. 38. Randomized Hash Lock Analysis <ul><li>+ Implementation requires hash and random number generator </li></ul><ul><ul><ul><li>Low-cost PRNG. </li></ul></ul></ul><ul><ul><ul><li>Physical randomness. </li></ul></ul></ul><ul><li>+ Randomized response prevents tracking. </li></ul><ul><li>- Inefficient brute force key look-up. </li></ul><ul><li>Hash is only guaranteed to be one-way. Might leak information about the ID. </li></ul><ul><li>(Essentially end up with a block cipher?) </li></ul>
    37. 39. Blocker Tags <ul><li>Juels, Rivest, Szydlo (2003). </li></ul><ul><li>Consumer Privacy Protecting Device: </li></ul><ul><ul><li>Hides your tag data from strangers. </li></ul></ul><ul><li>Users carry a “blocker tag” device. </li></ul><ul><li>Blocker tag injects itself into the tag’s anti-collision protocol. </li></ul><ul><li>Effectively spoofs non-existent tags. </li></ul><ul><li>(Only exists on paper.) </li></ul>
    38. 40. Other Work <ul><li>Efficient Implementations for RFID: </li></ul><ul><ul><li>Feldhofer, Dominikus, and Wolkerstorfer. </li></ul></ul><ul><ul><li>Gaubatz, Kaps, and Yüksel. </li></ul></ul><ul><li>Secure Protocols: </li></ul><ul><ul><li>Ari Juels. </li></ul></ul><ul><ul><li>Inoue and Yasuura </li></ul></ul><ul><ul><li>Gildas Avoine. </li></ul></ul><ul><li>Privacy Issues: </li></ul><ul><ul><li>Molnar and Wagner. </li></ul></ul><ul><ul><li>Henrici and Müller. </li></ul></ul>Limited Bibliography:
    39. 41. RFID Policy <ul><li>Policy can address a lot of privacy issues. </li></ul><ul><li>RSA Security is proposing a “privacy bit”: </li></ul><ul><ul><li>Sort of like a “do not disturb” sign. </li></ul></ul><ul><ul><li>Doesn’t stop someone from reading a tag. </li></ul></ul><ul><ul><li>More bits could encode various access policies </li></ul></ul><ul><li>Garfinkel has proposed an RFID Bill of Rights. </li></ul><ul><li>Other fair information practices proposed by EPIC, EFF, CASPIAN, etc. </li></ul>
    40. 42. Simson’s Bill of Rights <ul><li>The RFID Bill of Rights: </li></ul><ul><ul><li>The right to know whether products contain RFID tags. </li></ul></ul><ul><ul><li>The right to have RFID tags removed or deactivated when they purchase products. </li></ul></ul><ul><ul><li>The right to use RFID-enabled services without RFID tags. </li></ul></ul><ul><ul><li>The right to access an RFID tag’s stored data. </li></ul></ul><ul><ul><li>The right to know when, where and why the tags are being read. </li></ul></ul>
    41. 43. A New Idea: Humans and Tags <ul><li>Tags are dumb. But so are people. </li></ul><ul><li>Hopper and Blum have human-oriented identification protocols that you can do in your head. Linked off </li></ul><ul><li>Now adopting their protocol to RFID and securing it against stronger adversaries. </li></ul><ul><li>(Papers in progress.) </li></ul>
    42. 44. Questions?
    43. 45. Don’t forget to vote!