On inter-realm authentication in large distributed systems

Summary form only given. The authors define formally and rationalize a set of policy rules for the propagation of authentication trust across realm boundaries. The use of these rules helps limit global security exposures that ensue whenever an authentication service is compromised. The authors present the design of a simple, general protocol which satisfies these rules and shares the simplicity and generality goals of the protocol outlined by A. D. Birrell et al. (1986) and of the protocols implemented in Kerberos V (J. Kohl and C. Neuman, 1990) and in SPX (J. J. Tardo and K. Alagappan 1991)