Jan Krzysztof Miziołek

Parameterized Role-Based Access Control Policies for XML Documents

ABSTRACT Role-based access control policies (RBAC) are often used to provide access to fragments of static XML documents. Existing implementations of such RBACs often disseminate a single document encrypted with multiple cryptographic keys. However, most existing approaches are subject to role proliferation, especially in the case of large organizations where the number of defined roles may be several hundred. In such circumstances, correctly administering access control becomes much more difficult and error-prone. In this article, we present a novel approach to RBACs, which supports role parameterization to mitigate the potential of role proliferation. Our approach supports the association of specific user and/or session-specific credentials (i.e., parameters) with roles. We first define parameterized RBAC (PRABC), and then provide an algorithm for generating the minimal set of keys required to enforce a particular parameterized policy. We present another algorithm for efficiently encrypting an XML document in a single pass, using a technique that disguises the original structure of hidden subtrees. Finally, we include a key distribution algorithm that ensures each user receives only those keys that are needed for decrypting accessible fragments of the document. We analyze the complexity of our implementation and provide experiments to demonstrate its scalability.

Parallelization of an XML data compressor on multi-cores

Because of a growing interest in using XML for massive complex data there has been considerable research on designing XML compressors. This paper presents our research aimed at building parallel XML compressors, using Java and OpenMP (with C++). Our findings show that OpenMP is a preferred choice achieving better results than Java using a multi-core platform.

Permutation Based XML Compression

An XML document D often has a regular structure, i.e., it is composed of many similarly named and structured subtrees. Therefore, the entropy of a trees structuredness should be relatively low and thus the trees should be highly compressible by transforming them to an intermediate form. In general, this idea is used in permutation based XML-conscious compressors. An example of such a compressor is called XSAQCT, where the compressible form is called an annotated tree. While XSAQCT proved to be useful for various applications, it was never shown that it is a lossless compressor. This paper provides the formal background for the definition of an annotated tree, and a formal proof that the compression is lossless. It also shows properties of annotated trees that are useful for various applications, and discusses a measure of compressibility using this approach, followed by the experimental results showing compressibility of annotated trees.

Networked XML Compression by Encoding Pre-order Traversals

The advantages of the eXtensible Markup Language, XML, come at a cost, especially for huge datasets or when used on small mobile devices. Several known XML-conscious compressors used in real time environments compress data during data streaming. This paper presents a study of new real time algorithms that exploit local structural redundancies of pre-order traversals of an XML tree. These algorithms focus on reducing the overhead of streaming data while maintaining load balancing between the sender and receiver. Our algorithms have similar or better performance than existing algorithms, while emphasizing low memory and processing overheads.

Parallelization of Permuting XML Compressors

The verbose nature of XML results in overheads in storage and network transfers, which may be overcome by using parallel computing. This paper presents four permuting parallel XML compressors, based on an existing XML compressor, called XSAQCT. Tests were performed on multi-core machines using a test suite incorporating XML documents with various characteristics, and results were analyzed to find upper bounds given by Amdahl’s law, the actual speedup, and compression ratios.

Schema‐level access control policies for XML documents

Purpose – The purpose of this paper is to consider the secure publishing of XML documents, where a single copy of an XML document is disseminated and a stated role‐based access control policy (RBACP) is enforced via selective encryption. It describes a more efficient solution over previously proposed approaches, in which both policy specification and key generation are performed once, at the schema‐level. In lieu of the commonly used super‐encryption technique, in which nodes residing in the intersection of multiple roles are encrypted with multiple keys, it describes a new approach called multi‐encryption that guarantees each node is encrypted at most once.Design/methodology/approach – This paper describes two alternative algorithms for key generation and single‐pass algorithms for multi‐encrypting and decrypting a document. The solution typically results in a smaller number of keys being distributed to each user.Findings – The paper proves the correctness of the presented algorithms, and provides experi...