24-03-09 12:12
Age: 3 yrs

BY: DUC, DANG NGUYEN; LEE, HYUNROK; KIM, KWANGJO

It is well known that RFID is subject to various security threats, most notably tag cloning and tracking. To cope with these security threats, we need to implement cryptographic protocols on RFID tags. However, designing a cryptographic protocol is a difficult process. It is even more difficult when the design is restricted by the limited computational power of the targeted devices. Meanwhile, RFID tag is perhaps the device with least computational power due to a very tight price constraint of a RFID tag. Therefore, designing a secure yet lightweight cryptographic protocol for RFID tags is both challenging and tempting. There are two approaches in designing cryptographic protocols for low cost and low computational power devices: finding more efficient implementation of existing protocols and designing new lightweight protocols from ground-up. This paper is about the latter. A foundation for security of a cryptographic protocol is a hard computational problem. Intuitively speaking, a cryptographic protocol is said to be secure if breaking security is computationally equal to solving a hard problem. Popular hard problems for existing cryptographic protocols include integer factoring (IP), discrete logarithm (DLP) and Diffie-Hellman problem (DHP). In this whitepaper, we discuss the advantages of designing cryptographic protocols for RFID tags based on unconventional hard problems rather than IP, DLP or DHP. We show an example by presenting several lightweight cryptographic protocols based on a hard learning problem called Learning Parity with Noise problem (LNP).

<- Back to: Publications