Research On Anonymous Privacy Model And Technology In RFID Authentication System

Radio Frequency Identification (RFID) technology is a new wireless sensing technology, which uses radio waves to automatically identify things. RFID has the advantages such as identification speed, good environmental adaptability, and reusable. It is mainly used in the fields of identification and authentication, which is especially widely used in intelligent transportation, identification, logistics management, etc.Due to the exposure of wireless communication channel, the defects of interaction protocol, and the security risk of various components, the communication between the reader and tags can be eavesdropped or replayed by attackers, who can also carry out tracking attack on the tag. Hence, the RFID technology is suffering security and privacy issues. Moreover, with the expansion of the system, the number of tags is more and more bigger, which raised the following questions. On the one hand, the huge number of tags forced manufacturers to constantly drive down the cost of tag production, that lead to the shortage of computing power and storage resources of tags. So the tag can only perform some simple operations such as XOR, PRNG, Hash function, etc. but cannot support the complex encryption technology. The RFID system requirements, on the other hand, the reader can quickly complete identification and authentication of the tag, reduce time delay and ensure high efficiency of authentication. Obviously, under the condition of limited resources, how to realize the balance between privacy and authentication efficiency, becomes the key problem.In order to solve the problem, researchers have proposed a lot of privacy models and lightweight authentication protocols. For example, Hash-Lock protocol proposed by Sarma, the tree structure protocol proposed by Weis and the refresh privacy proposed by Li Lu et al. However, they are all have some problems in terms of security or authentication efficiency. Such as Hash-Lock protocol carries the risk of information leakage, and low efficiency of authentication. The tree structure protocol needs more storage resources, and it’s vulnerable to the compromised attack. Refresh privacy model is vulnerable to tracking attack, resulting in privacy loss of tag.This paper mainly studies how to solve the problem of security and privacy in the RFID systems under the premise of the low cost. Through in-depth analysis of the various components of the system, we can understand its characteristics and threats. To research and analysis the defects of the current privacy model and interaction protocol, we give optimization and solutions. In the face of new attack methods, we put forward a new privacy protection model and give the available authentication protocol. Meanwhile, we will pay more attention to the balance between privacy protection and authentication efficiency.The main work in this paper is as follows:(1) In RFID authentication system, we classify the threat of privacy with the demand of privacy protection, and give the solution by analysis of problems existing in the privacy model. Strong privacy has the strict constrains on the output of tags, such as the randomization and unpredictability. It’s leading to reduce authentication efficiency. Refresh is a weak privacy model, as the name suggests, it is mainly based on the refresh mechanism. Refresh loosens the strict constrains on the output of tags, such as the randomization and unpredictability. Refresh allows a tag to contain a temporally constant field in the output. So refresh will use this information to improve the efficiency of authentication. But the attacker can track the tag by binding it with the information which is easily obtain by eavesdropping or intercepting. So the information of the tag’s location will be leaked. Although, the attacker didn’t know the tag’s key but it is still a serious threat to the user’s personal privacy.(2) We propose a new privacy protection model of RFID system based on k-anonymous. The k-anonymous is a classic method for privacy protection of data published through structuring the k similar records in the data, so the attacker cannot distinguish between k records. In particular,k-anonymous is a way to generalize a certain attribute value of the data by concept hierarchy generalization, to ensure each attribute value corresponding to at least k records. Even though an attacker gains an attribute value in other ways, he cannot distinguish between k records, unable to locate a specific record, so the user’s privacy is protected. In the new privacy model, we use the ideas of k-anonymous to deal with tags. All tags in the system are randomly divided into several groups, and each group contains at least k tags which use the same quasi-identifier. Quasi identifier between the groups are not the same and independent of each other. Tag’s response information will contain this quasi identifier in the form of plaintext, but there is no the security issues like in the refresh model. Because, even if the attacker obtains quasi identifier, he cannot distinguish tags in the same group. He also cannot track the tag. So the new model can ensure the tag’s location privacy be safe. At the same time, the quasi identifier can be used to speed up the authentication efficiency by reader.(3) We further propose a new light-weight privacy preserving authentication protocol, k-PPA, which can guarantee the privacy based on the k-anonymous privacy model and have high authentication efficiency, simultaneously. The k-PPA protocol can resist eavesdropping attacks, replay attacks, forgery attacks, and location tracking attacks. Compared with similar protocol,k-PPA protocol can achieve a better balance between privacy and authentication efficiency.