We present representations and algorithms for the implementation of RTC, a role-based trust management language, and announce an open-source implementation available to the public. We also design and perform large-scale performance tests on policies closely modeled after possible applications of RT in the real world. These tests aim to determine the viability of RT as an authorization solution for large and potentially complex policies in a decentralized environment; the results of the tests are analyzed to identify what policy characteristics most strongly affect the performance of RT and develop strategies to achieve the rapid response times required in real-world authorization systems.