Description
Performance considerations played an important role in the design of the Authenticated Datagram Protocol (ADP), a subtransport-level host-to-host datagram protocol that contains cryptographic mechanisms for end-to-end authentication and, optionally, privacy of messages. Several performance-motivated features were introduced into ADP. This paper describes the first phase of a measurement-based study of ADP intended to determine the actual effects of each of those features on the protocol's performance. The experiments were trace-driven, and took place between two workstations in a laboratory setting. The results in every case demonstrated the usefulness of the features in question, but not always for the reasons that motivated their introduction into the design.