Performance analysis of identification codes

Abstract

In this paper, we analyze the construction of identification codes. Identification codes are based on the question: "Is the message I have just received the one I am interested in?", as opposed to Shannon's transmission, where the receiver is interested in not only one, but any, message. The advantage of identification is that it allows rates growing double exponentially in the blocklength at the cost of not being able to decode every message, which might be beneficial in certain applications. We focus on a special identification code construction based on two concatenated Reed-Solomon codes and have a closer look at its implementation, analyzing the trade-offs of identification with respect to transmission and the trade-offs introduced by the computational cost of identification codes.