Constructing practical and provably secure group signature schemes has been a very active research topic in recent years. A group signature can be viewed as a digital signature with certain extra properties. Notably, anyone can verify that a signature is generated by a legitimate group member, while the actual signer can only be identified (and linked) by a designated entity called a group manager. Currently, the most efficient group signature scheme available is due to Camenisch and Lysyanskaya \cite{CL02}. It is obtained by integrating a novel dynamic accumulator with the scheme by Ateniese, et al. \cite{ACJT00}. In this paper, we construct a dynamic accumulator that accumulates \emph{composites}, as opposed to previous accumulators that accumulated \emph{primes}. We also present an efficient method for proving knowledge of factorization of a committed value. Based on these (and other) techniques we design a novel provably secure group signature scheme. It operates in the \emph{common auxiliary string} model and offers two important benefits: 1) the {\sf Join} process is very efficient: a new member computes only a single exponentiation, and 2) the (unoptimized) cost of generating a group signature is 17 exponentiations which is appreciably less than the state-of-the-art.

Biometric personal authentication systems have a common problem --- the biological information can easily be stolen by other individuals. In line with the process of the activities for the international standardization of the biometric system, this paper proposes a typical way to embed biological information, whatever its kind, into cryptographic keys as a measure for privacy protection and against unauthorized use. We believe that our proposal presents the following advantages: the improvement of protecting the privacy of biological information, economical effectiveness resulting from the practical use of the infrastructure of Public Key Infrastructure (PKI) as a biological information database, and humanity given to a man-machine interface by embedding an individual's biological information into a public key, an important element of the system. This paper also proposes how to build up a practical personal authentication system through the method proposed.

We consider how to avoid unsolicited e-mail -- so called spam -- in a stronger adversarial model than has previously been considered. Our primary concern is the proposal of an architecture and of protocols preventing against successful spamming attacks launched by a strong attacker. This attacker is assumed to control the communication media and to be capable of corrupting large numbers of protocol participants. Additionally, the same architecture can be used as a basis to support message integrity and privacy, though this is not a primary goal of our work. This results in a simple and efficient solution that is largely backwards-compatible, and which addresses many of the concerns surrounding e-mail communication.

In distributed networks, a target party $T$ could be a person never meet with a source party $S$, therefore $S$ may not hold any prior evaluation of trustworthiness of $T$. To get permit to access $S$, $T$ should be somewhat trusted by $S$. Consequently, we should study the approach to evaluate trustworthiness of $T$. To attack the problem, we view individual participant in distributed networks as a node of a delegation graph $G$ and map a delegation path from target party $T$ to source party $S$ in networks into an edge in the correspondent transitive closure of graph $G$. Based on the transitive closure property of the graph $G$, we decompose the problem to three related questions below: -how to evaluate trustworthiness of participants in an edge? -how to compute trustworthiness of participants in a path? -how to evaluate the trustworthiness of a target participant in a transitive closure graph? We attack the above three questions by first computing trustworthiness of participants in distributed and authenticated channel. Then we present a practical approach to evaluate trustworthiness by removing the assumption of the authenticated channel in distributed networks.

This paper examines mechanical lock security from the perspective of computer science and cryptology. We focus on new and practical attacks for amplifying rights in mechanical pin tumbler locks. Given access to a single master-keyed lock and its associated key, a procedure is given that allows discovery and creation of a working master key for the system. No special skill or equipment, beyond a small number of blank keys and a metal file, is required, and the attacker need engage in no suspicious behavior at the lock's location. Countermeasures are also described that may provide limited protection under certain circumstances. We conclude with directions for research in this area and the suggestion that mechanical locks are worthy objects for study and scrutiny.

We propose a new technique for making mix nets robust, called randomized partial checking (RPC). The basic idea is that rather than providing a proof of completely correct operation, each server provides strong evidence of its correct operation by revealing a pseudo-randomly selected subset of its input/output relations. Randomized partial checking is exceptionally efficient compared to previous proposals for providing robustness; the evidence provided at each layer is shorter than the output of that layer, and producing the evidence is easier than doing the mixing. It works with mix nets based on any encryption scheme (i.e., on public-key alone, and on hybrid schemes using public-key/symmetric-key combinations). It also works both with Chaumian mix nets where the messages are successively encrypted with each servers' key, and with mix nets based on a single public key with randomized re-encryption at each layer. Randomized partial checking is particularly well suited for voting systems, as it ensures voter privacy and provides assurance of correct operation. Voter privacy is ensured (either probabilistically or cryptographically) with appropriate design and parameter selection. Unlike previous work, our work provides voter privacy as a global property of the mix net rather than as a property ensured by a single honest server. RPC-based mix nets also provide very high assurance of a correct election result, since a corrupt server is very likely to be caught if it attempts to tamper with even a couple of ballots.

This paper describes malicious applets that use Java's sophisticated graphic features to rectify the browser's padlock area and cover the address bar with a false https domain name. The attack was successfully tested on Netscape's Navigator and Microsoft's Internet Explorer; we consequently recommend to neutralize Java whenever funds or private data transit via these browsers and patch the flaw in the coming releases. The degree of novelty of our attack is unclear since similar (yet non-identical) results can be achieved by spoofing as described by Felten; however our scenario is much simpler to mount as it only demands the inclusion of an applet in the attacker's web page. In any case, we believe that the technical dissection of our malicious Java code has an illustrative value in itself.