SFSDA: Secure and Flexible Subset Data Aggregation with Fault Tolerance for Smart Grid

Smart grid (SG) brings convenience to users while facing great chal-lenges in protecting personal private data. Data aggregation plays a key role in protecting personal privacy by aggregating all personal data into a single value, preventing the leakage of personal data while ensuring its availability. Recently, a flexible subset data aggregation (FSDA) scheme based on the Pail-lier homomorphic encryption was first proposed by Zhang et al. Their scheme can dynamically adjust the size of each subset and obtain the aggregated data in the corresponding subset. In this paper, firstly, an efficient attack with both theorems proving and experimentative verification is launched. We find that in a specific scenario where the encrypted data constructed by a smart meter (SM) exceeds the size of one Paillier ciphertext, the malicious fog node (FN) may use the received ciphertext to obtain the reading of the SM. Secondly, to avoid the possibility of privacy disclosure under certain circumstances, additional hash functions are added to the individual encryption process. In addition, fault tolerance is very important to aggregation schemes in practical scenarios. In most of the current schemes, once some SMs failed, then they will not work. As far as we know, there is no multi-subset aggregation scheme both supports flexible subset data aggregation and fault tolerance. Finally, we construct the first secure flexible subset data aggregation (SFSDA) scheme with fault tolerance by combining the fault tolerance method with the flexible multi-subset aggregation, where FN enables the control server (CS) to finally decrypt the aggregated ciphertext by recovering equivalent ciphertexts when some SMs fail to submit their ciphertexts. Experiments show that our SFSDA scheme keeps the efficiency in implementing a flexible multi -subset aggregation function, and only has a small delay in implementing fault -tolerant data aggregation.