دانلود رایگان مقاله اندازه گیری هوشمند کارآمد بر اساس رمزنگاری همریختی

abstract

Smart meters send fine-grained client electricity consumption readings to suppliers. Although this presents advantages for both entities, it results in a serious loss of privacy for customers. We present a monitoringpurpose system that preserves customers’ privacy by homomorphically aggregating the consumptions of all n members of a neighborhood. The proposal has an efficient linear O(n) communication cost and is proven to preserve customers’ privacy even in the presence of a corrupted substation and some malicious smart meters. It requires neither secure communication channels nor a trusted third party (except for issuing public-key certificates). Computation on the smart meters is limited to modular exponentiations. These favorable properties come at the expense of increased computation cost on the electricity suppliers’ side. We show that the computation is easily feasible for realistic parameter choices.

5. Conclusions

In this paper, we have presented an efficient privacy-preserving system for reporting the consumption of a neighborhood of n smart meters. By homomorphically adding all n consumptions, the existing link between customers and their individual consumption values is broken. In this way, detailed information can be sent without leaking individual personal data. Our approach does not require a trusted third party (except a certification authority), and has a linear O(n) communication complexity. In contrast to some other approaches [12,18], our solution does not require communication among smart meters, but only with the electricity supplier (represented by the substation). The individual reading of a smart meter has been proven to be kept private even assuming a corrupted substation. Malicious smart meters are also acceptable—the precise reading of one smart meter is only revealed if all others in the group are malicious. The hardest computations to be performed by the smart meters are a few modular exponentiations per round, which are easily feasible despite possible resource limitations. While computation overhead on the substation is larger, we have shown that this does not pose a problem in a realistic setting—even without any optimizations.