Max Kanovich Made a Report "Can we mitigate the attacks on Distance-Bounding Protocols by using challenge-response rounds repeatedly?"

April 21, Max Kanovich, professor of computer science at Department of Computer Science and Information Systems, University of London, visiting professor of Faculty of Computer Science,School of Data Analysis and Artificial Intelligence, made a report at the session of our seminar.

 

The core of  Distance-Bounding Protocols  is a kind of the challenge-response sub-protocol  to identify the participants, Prover and Verifier.

 The aim of the talk is twofold:

 1. First,

 for a typical challenge-response protocol, presented here, we give a full probabilistic analysis of an attack ``between ticks'', newly established by Max Kanovich, Tajana Ban Kirigin,   Vivek Nigam, Andre Scedrov, and Carolyn Talcott.Their attack is based on the discrepancy between the  'observable' challenge-response time interval  and the 'actual' challenge-response time interval;  the dicrepancy is caused by inconsistency  between the continuous time in nature and  the discrete time within the computer clock.

 2. Secondly,

 we challenge a kind of a general belief that  within Distance-Bounding Protocols  Verifier can improve their performance  by means of collecting statistics in a series  of independent challenge-response rounds  aiming to observe an  ``acceptance challenge-response event''  in the majority of the rounds.

The novelty of our approach is that here we get quite surprising results to support such a claim as well as  to disprove it.

Namely, we show that in the case where  Verifier decides to grant the access by the simple majority, the effect of the  repeated challenge-response rounds  can mitigate the attack but only for  the specific values of the probability  of the erroneous decision in one round.

Whereas in the case where Verifier  decides to grant the access by the large  majority (that is, with gaining a specified  level of support which is greater, say two thirds,  than the threshold of one half used for simple  majority) the idea of repeated challenge-response  rounds works perfectly well for our protocol.  In particular, having observed the ``acceptance challenge-response events''  in the two-thirds majority of rounds,  Verifier can establish the desired upper  bounds for the 'actual' challenge-response  time interval but only with the  high probability.