On Ultralightweight RFID Authentication Protocols

A recent research trend, motivated by the massive deployment of RFID technology, looks at cryptographic protocols for securing communication between entities in which some of the parties have very limited computing capabilities. In this paper, we focus our attention on SASI, a new RFID authentication protocol, designed for providing Strong Authentication and Strong Integrity. SASI is a good representative of a family of RFID authentication protocols, referred to as Ultralightweight RFID authentication protocols. These protocols, suitable for passive Tags with limited computational power and storage, involve simple bitwise operations such as and, or, exclusive or, modular addition, and cyclic shift operations. They are efficient, fit the hardware constraints, and can be seen as an example of the above research trend. However, the main concern is the real security of these protocols, which are often supported only by apparently reasonable and intuitive arguments. The contribution we provide with this work is the following: we start by showing some weaknesses in the SASI protocol, and then, we describe how such weaknesses, through a sequence of simple steps, can be used to compute in an efficient way all secret data used for the authentication process. Specifically, we describe three attacks: 1) a desynchronization attack, through which an adversary can break the synchronization between the RFID Reader and the Tag; 2) an identity disclosure attack, through which an adversary can compute the identity of the Tag; and 3) a full disclosure attack, which enables an adversary to retrieve all secret data stored in the Tag. Then, we present some experimental results, obtained by running several tests on an implementation of the protocol, in order to evaluate the performance of the proposed attacks, which confirm that the attacks are effective and efficient. It comes out that an active adversary by interacting with a Tag more or less three hundred times, makes the authentication protocol completely useless. Finally, we close the paper with some observations. The cryptoanalysis of SASI gets some new light on the ultralightweight approach, and can also serve as a warning to researchers working on the field and tempted to apply these techniques. Indeed, the results of this work, rise serious questions regarding the limits of the ultralightweight family of protocols, and on the benefits of these ad hoc protocol design strategies and informal security analysis.