Perfectly secure message transmission over partially synchronous networks

In a distributed network, we consider two special nodes called the sender S and the receiver R that are connected by n node-disjoint (except for S and R) bi-directional wires. Out of these n wires, the adversary can control at most t wires (of its choice) in Byzantine fashion. In this setting, our goal is to design a message transmission protocol Π that assures the following two conditions hold: (1) by the end of the protocol Π, R gets the correct message m transmitted by S without any error (perfect reliability), and (2) the adversary learns no information about m, whatsoever, in information theoretic sense (perfect secrecy). Protocols that satisfy these two conditions are known as the Perfectly Secure Message Transmission (PSMT) protocols. However, out of the n wires that exist, if some number of wires say ns, fortunately, happen to be synchronous (serendipitous synchrony) then we ask under what conditions do PSMT protocols tolerating t-Byzantine faults exist. In the literature, it is known that, if either ns > 2t or n > 3t then PSMT protocols trivially exist. Therefore, we consider the case where we have at most 2t synchronous wires (i.e., ns ≤ 2t) and at most 3t wires overall (i.e., n ≤ 3t). Interestingly, we prove that in this case, no PSMT protocol exists. This concludes that, in designing PSMT protocols (tolerating the given fixed number of faults), either (serendipitous) synchronous wires alone are sufficient or we get absolutely no extra advantage of a wire being synchronous over asynchronous.

[1]  K. Srinathan,et al.  Secure message transmission in asynchronous networks , 2011, J. Parallel Distributed Comput..

[2]  K. Srinathan,et al.  Unconditionally reliable and secure message transmission in undirected synchronous networks: possibility, feasibility and optimality , 2010, Int. J. Appl. Cryptogr..

[3]  K. Srinathan,et al.  On perfectly secret message transmission in digraphs tolerating dual failures , 2016, ICDCN.

[4]  Yongge Wang,et al.  Perfectly Secure Message Transmission Revisited , 2008, IEEE Trans. Inf. Theory.

[5]  K. Srinathan,et al.  On the trade-off between network connectivity, round complexity, and communication complexity of reliable message transmission , 2012, JACM.

[6]  K. Srinathan,et al.  On perfectly secure communication over arbitrary networks , 2002, PODC '02.

[7]  Hosame Abu-Amara,et al.  Perfectly secure message transmission in asynchronous networks , 1995, Proceedings.Seventh IEEE Symposium on Parallel and Distributed Processing.

[8]  Seif Haridi,et al.  Distributed Algorithms , 1992, Lecture Notes in Computer Science.

[9]  Kannan Srinathan,et al.  On the Price of Proactivizing Round-Optimal Perfectly Secret Message Transmission , 2018, IEEE Transactions on Information Theory.

[10]  K. Srinathan,et al.  Asynchronous Secure Communication Tolerating Mixed Adversaries , 2002, ASIACRYPT.

[11]  Matthias Fitzi,et al.  Towards Optimal and Efficient Perfectly Secure Message Transmission , 2007, TCC.

[12]  Moti Yung,et al.  Perfectly secure message transmission , 1993, JACM.

[13]  C. Pandu Rangan,et al.  Constant phase efficient protocols for secure message transmission in directed networks , 2007, PODC '07.

[14]  K. Srinathan,et al.  On minimal connectivity requirements for secure message transmission in directed networks , 2018, Inf. Process. Lett..

[15]  K. Srinathan,et al.  On Minimal Connectivity Requirement for Secure Message Transmission in Asynchronous Networks , 2009, ICDCN.

[16]  K. Srinathan,et al.  Interplay between (im)perfectness, synchrony and connectivity: The case of reliable message transmission , 2013, Theor. Comput. Sci..

[17]  K. Srinathan,et al.  Optimal Perfectly Secure Message Transmission , 2004, CRYPTO.

[18]  Matthew K. Franklin,et al.  Secure Communication in Minimal Connectivity Models , 1998, Journal of Cryptology.