Further Reflections on RFID Security

Adi Shamir once again gave security researchers something to reflect on with his announcement at RSA Conference 2006 of a new side-channel attack on the RFID "kill" function. His attack, which has been implemented in the lab, analyzes the radio energy returned by a standard RFID tag to determine the tag's secret "kill code." Once the attacking equipment -- basically an RFID reader with an extra-sensitive antenna -- has recovered the kill code, it can then deactivate the tag. A typical implementation of a comparison function sets a flag when it finds the first bit that doesn't match. Setting a flag takes energy. If a passive RFID tag implements the comparison function this way, the energy it uses to set the tag won't be reflected back to the reader. A rogue reader can thereby identify one of the bits of the kill code that's incorrect. It can correct this bit and try again until all the bits are correct. My colleague Ari Juels has expounded on the attack and its implications in an article published earlier this week. Shamir suggested a cell phone as the bearer of the attack. I'm envisioning a modified version of one of those robotic vacuum cleaners, roaming a warehouse and sucking kill codes out of all the tags in the inventory. It's clear now in practice that as RFID security pioneers Sanjay Sarma, Stephen Weis and Daniel Engels suggested at CHES 2002, "Being contactless and passively powered may make RFID devices especially susceptible to fault induction or power analysis attacks." ^[1] It shouldn't be too hard to avoid the attack -- just set one flag when a bit doesn't match, and set a different flag when it does. If setting either flag takes the same amount of energy, it will be much harder for a reader to tell what's going on. Still, it's a good time for designers of RFID systems to reflect on these issues -- before the rogue cell phones and the roaming readers of the future are the ones doing the reflection.

^[1] Sanjay E. Sarma, Stephen A. Weis, and Daniel W. Engels. RFID Systems and Security and Privacy Implications. In Cryptographic Hardware in Embedded Systems, volume 2523 of Lecture Notes in Computer Science, pages 454-470, 2002. Available via http://theory.lcs.mit.edu/~sweis/biblio.html.

Post A Comment

Your Name
Your Email	Publish email?: Yes No
Your Blog
Subject
Comment
Verification Word