A variety of emerging microelectronics applications target portable systems with tight constraints on the related metrics of power, form factor, and longevity. For many of these applications, there are severe constraints on the energy consumption for the electronics in the system. In particular, passive RFID tags rely on power received from readers so low power consumption is necessary to enable long-range reads. Nevertheless, these devices are already widely used for applications with serious security and privacy requirements such as key cards, public transportation tokens, and implantable medical devices. Current design approaches and standard cryptographic primitives fail to satisfy the needs for these systems. As a result, current implementations either resort to "security-by-obscurity" and ad hoc solutions that fail to provide adequate security and are frequently broken in practice.
This project is developing a comprehensive approach to analyzing, designing, and implementing security and privacy for severely resource-constrained devices such as widely deployed RFID systems. The goal is to enable designers to create secure, cost-effective, large-scale RFID-deployments by combining primitives and protocols from a library with known properties and to implement those designs in a principled and efficient manner. The research follows an interdisciplinary approach by bringing together experts in low-energy integrated circuit design, design automation and embedded systems, system and network security, and cryptography. In addition to the research, the project includes a modular RFID security lab course and teaching outreach courses using cryptography and RFID systems to excite middle school students about pursuing science and engineering.
Low-powered devices that provide wireless identification and authentication, including RFID tags, are everywhere today. They are embedded in credit cards, public transit tokens, and access cards that millions of us carry every day. The signals these devices send can be intercepted and attackers may be able to use intercepted messages to clone the devices or track the owners of these devices. Our research in this area developed from studying the security and privacy of the most widely deployed RFID tags, the MIFARE chips that are used in billions devices including car key immobilizers, door access cards, and public transportation keys (e.g., Boston’s MBTA). By analyzing the circuits in these chips, we identified several serious security weaknesses that would enable an eavesdropper to clone a card or track individuals based on the cards they carry and worked with vendors to mitigate these vulnerabilities and improve the cryptosystems used in subsequent devices. To address the more fundamental problems, we developed a model for privacy that allows engineers to measure the tradeoffs between privacy and cost in a meaningful way. We designed a new cryptosystem that provides additional points in the design space and employs a new type of cryptography to provide potentially high security without requiring any operations more expensive than addition. We tested implementations of our design in low-power circuits, and demonstrated the feasibility of our approach. Our work opens up new strategies for designing circuits that enable engineers to trade-off fidelity and efficiency, which can be useful for a wide range of applications, including the particular kinds of cryptosystems we employ in our RFID designs.
