Server-aided secret computation (SASC) protocols were proposed as methods to allow a not-so-powerful computer (a client) to compute a function efficiently with the aid of a powerful computer (a server) without revealing the client's secret to the server. A smart card system is considered to be the most promising application for SASC protocols. Matsumoto et al. proposed a class of protocols, called S2, which is suitable for RSA secret computation. It includes many variations, such as binary S2, non-binary S2, and KS schemes. Although it is expected that the non-binary S2 will be the most efficient scheme, performances for these protocols have not been fully analyzed, except for the KS scheme. This paper discusses how to determine appropriate parameters for non-binary S2. Performance was generally analyzed by using normalized variables. It is shown that a client can compute a signature 50 times faster, with a 105 times more powerful server's aid, than in the case without a server, provided that communication cost can be ignored. The SASC protocol performance is also discussed under the practical condition that a server is an 8-bit CPU and that the communication rate is 9.6 kbps. The client can sign a message in about 10 seconds, with the aid of a 32-bit CPU. This analysis coincides well with the experimental result. It is estimated that a client with a 16-bit CPU will accomplish a less that 2 second processing time with a 4.6 kbps server.
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copy
Shin-ichi KAWAMURA, Atsushi SHIMBO, "Performance Analysis of Server-Aided Secret Computation Protocols for the RSA Cryptosystem" in IEICE TRANSACTIONS on transactions,
vol. E73-E, no. 7, pp. 1073-1080, July 1990, doi: .
Abstract: Server-aided secret computation (SASC) protocols were proposed as methods to allow a not-so-powerful computer (a client) to compute a function efficiently with the aid of a powerful computer (a server) without revealing the client's secret to the server. A smart card system is considered to be the most promising application for SASC protocols. Matsumoto et al. proposed a class of protocols, called S2, which is suitable for RSA secret computation. It includes many variations, such as binary S2, non-binary S2, and KS schemes. Although it is expected that the non-binary S2 will be the most efficient scheme, performances for these protocols have not been fully analyzed, except for the KS scheme. This paper discusses how to determine appropriate parameters for non-binary S2. Performance was generally analyzed by using normalized variables. It is shown that a client can compute a signature 50 times faster, with a 105 times more powerful server's aid, than in the case without a server, provided that communication cost can be ignored. The SASC protocol performance is also discussed under the practical condition that a server is an 8-bit CPU and that the communication rate is 9.6 kbps. The client can sign a message in about 10 seconds, with the aid of a 32-bit CPU. This analysis coincides well with the experimental result. It is estimated that a client with a 16-bit CPU will accomplish a less that 2 second processing time with a 4.6 kbps server.
URL: https://globals.ieice.org/en_transactions/transactions/10.1587/e73-e_7_1073/_p
Copy
@ARTICLE{e73-e_7_1073,
author={Shin-ichi KAWAMURA, Atsushi SHIMBO, },
journal={IEICE TRANSACTIONS on transactions},
title={Performance Analysis of Server-Aided Secret Computation Protocols for the RSA Cryptosystem},
year={1990},
volume={E73-E},
number={7},
pages={1073-1080},
abstract={Server-aided secret computation (SASC) protocols were proposed as methods to allow a not-so-powerful computer (a client) to compute a function efficiently with the aid of a powerful computer (a server) without revealing the client's secret to the server. A smart card system is considered to be the most promising application for SASC protocols. Matsumoto et al. proposed a class of protocols, called S2, which is suitable for RSA secret computation. It includes many variations, such as binary S2, non-binary S2, and KS schemes. Although it is expected that the non-binary S2 will be the most efficient scheme, performances for these protocols have not been fully analyzed, except for the KS scheme. This paper discusses how to determine appropriate parameters for non-binary S2. Performance was generally analyzed by using normalized variables. It is shown that a client can compute a signature 50 times faster, with a 105 times more powerful server's aid, than in the case without a server, provided that communication cost can be ignored. The SASC protocol performance is also discussed under the practical condition that a server is an 8-bit CPU and that the communication rate is 9.6 kbps. The client can sign a message in about 10 seconds, with the aid of a 32-bit CPU. This analysis coincides well with the experimental result. It is estimated that a client with a 16-bit CPU will accomplish a less that 2 second processing time with a 4.6 kbps server.},
keywords={},
doi={},
ISSN={},
month={July},}
Copy
TY - JOUR
TI - Performance Analysis of Server-Aided Secret Computation Protocols for the RSA Cryptosystem
T2 - IEICE TRANSACTIONS on transactions
SP - 1073
EP - 1080
AU - Shin-ichi KAWAMURA
AU - Atsushi SHIMBO
PY - 1990
DO -
JO - IEICE TRANSACTIONS on transactions
SN -
VL - E73-E
IS - 7
JA - IEICE TRANSACTIONS on transactions
Y1 - July 1990
AB - Server-aided secret computation (SASC) protocols were proposed as methods to allow a not-so-powerful computer (a client) to compute a function efficiently with the aid of a powerful computer (a server) without revealing the client's secret to the server. A smart card system is considered to be the most promising application for SASC protocols. Matsumoto et al. proposed a class of protocols, called S2, which is suitable for RSA secret computation. It includes many variations, such as binary S2, non-binary S2, and KS schemes. Although it is expected that the non-binary S2 will be the most efficient scheme, performances for these protocols have not been fully analyzed, except for the KS scheme. This paper discusses how to determine appropriate parameters for non-binary S2. Performance was generally analyzed by using normalized variables. It is shown that a client can compute a signature 50 times faster, with a 105 times more powerful server's aid, than in the case without a server, provided that communication cost can be ignored. The SASC protocol performance is also discussed under the practical condition that a server is an 8-bit CPU and that the communication rate is 9.6 kbps. The client can sign a message in about 10 seconds, with the aid of a 32-bit CPU. This analysis coincides well with the experimental result. It is estimated that a client with a 16-bit CPU will accomplish a less that 2 second processing time with a 4.6 kbps server.
ER -