Impossibility of secure cloud quantum computing for classical client
Tomoyuki Morimae, Takeshi Koshiba · arXiv (Cornell University) · 2014
The first generation quantum computer will be implemented in the cloud style, since only few groups will be able to access such an expensive and high-maintenance machine. How the privacy of the can be protected in such a cloud quantum computing? It was theoretically shown [A. Broadbent, J. F. Fitzsimons, and E. Kashefi, Proceedings of the 50th Annual IEEE Symposium on Foundation of Computer Science, 517 (2009)], and experimentally demonstrated [S. Barz, E. Kashefi, A. Broadbent, J. F. Fitzsimons, A. Zeilinger, and P. Walther, Science {\bf335}, 303 (2012)] that a who can generate randomly-rotated single qubit states can delegate her quantum computing to a remote quantum server without leaking any privacy. The generation of a single qubit state is not too much burden for the client, and therefore we can say that almost classical client can enjoy the secure cloud quantum computing. However, isn't is possible to realize a secure cloud quantum computing for a who is completely free from any quantum technology? Here we show that perfectly-secure cloud quantum computing is impossible for a completely classical unless classical computing can simulate quantum computing, or a breakthrough is brought in classical cryptography.