A bound of deception capability in multiuser computer networks

An expression is developed to quantify the deception capability of an attacker in a multiuser computer network. A communication ends up with deception when any one or both of the communicating parties are deceived. The word deception is used to mean either the act of spoofing on an established association (i.e., a bidirectional data link), the process of setting up an unauthorized association, or both. In general, an attacker can use several different procedures to cheat a legitimate user of a data communication network. A subset of those procedures is examined to formulate an expression for the probability of a successful attack. This formulation is intended to be helpful for the design of robust message structures and fault-tolerant link control protocols for computer communication networks. The tradeoffs involved in achieving a desired level of authenticity of communicated messages are discussed.