In EUROCRYPT 2011, Satoshi Obana presented two SSCI schemes, which can identify up to t < k / 2 cheaters. However, the schemes require |V_i| \approx (n (t+1) 2^{3t-1} |S|) / \epsilon and |V_i| \approx ((n t 2^{3t})^2 |S|) / (\epsilon^2)$ respectively. Moreover, both the schemes are computationally inefficient, as they require to perform exponential computation in general. So comparing our scheme with the schemes of Obana, we find that not only our scheme is computationally efficient, but in our scheme the share size is significantly smaller than that of Obana. Thus our scheme solves one of the open problems left by Obana, urging to design efficient SSCI scheme with t < k/2.
In CRYPT0 1995, Kurosawa, Obana and Ogata have shown that in any SSCI scheme, |V_i| \geq (|S| - 1) / (\epsilon) + 1. Though our proposed scheme does not exactly matches this bound, we show that our scheme {\it asymptotically} satisfies the above bound. To the best of our knowledge, our scheme is the best SSCI scheme, capable of identifying the maximum number of cheaters.
Category / Keywords: cryptographic protocols / Date: received 17 Jun 2011 Contact author: partho31 at gmail com Available format(s): PDF | BibTeX Citation Version: 20110622:200815 (All versions of this report) Short URL: ia.cr/2011/330 Discussion forum: Show discussion | Start new discussion