Improved topology assumptions for threshold cryptography in mobile ad hoc networks

Mobile Ad Hoc Networks (MANET), due to their lack of physical infrastructures or centralized authorities, pose a number of security challenges to a protocol designer. In particular, several typical application scenarios demand the design of protocols that cannot base their security on the existence of trusted parties or setup information, but rather need to leverage uniquely on assumptions limiting the corrupting power of the adversaries. This naturally defines security design and analysis paradigms similar to those of the Threshold Cryptography area, where it is typically assumed that an adversary can corrupt up to a limited amount of entities or resources. Therefore a secure realization of primitives from Threshold Cryptography in MANET promises to be applicable to several MANET protocols.Recently, in [10], we started the analysis of Threshold Cryptography solutions over MANET, by focusing on the problem of extending to these networks known efficient threshold signature schemes for wired networks. In particular, we noted a major design difficulty due to the lack of full network connectivity that significantly constrained the network topology assumptions under which a MANET threshold signature scheme can be proved secure. In this paper we continue our investigation and present a new MANET threshold signature scheme that is secure under significantly improved topology assumptions. Surprisingly, we break through an apparent barrier due to well-known results from the Distributed Computing area.