An efficient density peak cluster algorithm for improving policy evaluation performance

Abstract

Abstract In recent years, the XACML (eXtensible Access Control Markup Language) is widely used in a variety of research fields, especially in access control. However, when policy sets defined by the XACML become large and complex, the policy evaluation time increases significantly. In order to improve policy evaluation performance, we propose an optimization algorithm based on the DPCA (Density Peak Cluster Algorithm) to improve the clustering effect on large-scale complex policy sets. Combined with this algorithm, an efficient policy evaluation engine, named DPEngine, is proposed to speed up policy matching and reduce the policy evaluation time. We compare the policy evaluation time of DPEngine with the Sun PDP, HPEngine, XEngine and SBA-XACML. The experiment results show that (1) when the number of requests reaches 10,000, the DPEngine evaluation time on a large-scale policy set with 100,000 rules is approximately 2.23%, 3.47%, 3.67% and 4.06% of that of the Sun PDP, HPEngine, XEngine and SBA-XACML, respectively and (2) as the number of requests increases, the DPEngine evaluation time grows linearly. Compared with other policy evaluation engines, the DPEngine has the advantages of efficiency and stability.