Analyzing Protocol Security Through Information-Flow Control

Abstract

Security protocols are essential for establishing trust in electronic transactions over open networks. Currently used languages/logics for protocol specifications do not facilitate/force the designer to make explicit goals, intentional assumptions or the preceding history across interactions among the stakeholders. This has resulted in gaps in specifications which in turn have led to problems such as: (i) inefficient/non-optimal protocol designs, (ii) incompatible theoretical attacks discovered by analyzers due to different threat models and (iii) faulty or insecure implementations due to insufficient guidelines for the implementer. We have recently developed the readers-writers flow model (RWFM) that has several benefits, including simple and intuitive labels. In this paper, we demonstrate that the problem of incomplete protocol specification can be overcome by enriching them with labels from RWFM, which make explicit the assumptions and goals at each stage of the protocol. In particular, we use readers and writers as labels for data objects and roles for tracking information flows in a protocol that makes explicit the construction of new messages from components of previous messages and also the knowledge of roles at various stages. We illustrate our approach and demonstrate its advantages in comparison to prominent specification languages in the literature by using the example of Needham-Schroeder public key protocol. Further, we argue how the proposed approach leads to a robust protocol specification language including security/cryptographic protocols that shall be of immense aid to the designer, user and the implementer of protocols.

N.V. Narendra Kumar—The work was carried out with support from ISRDC (Information Security Research and Development Center), a project sponsored by MeitY, GoI.