Secure Aggregation Scheme Research Articles

VCSA: Verifiable and collusion-resistant secure aggregation for federated learning using symmetric homomorphic encryption

As a novel distributed learning framework for protecting personal data privacy, federated learning, (FL) has attained widespread attention through sharing gradients among users without collecting their data. However, an untrusted cloud server may infer users’ individual information from gradients and global model. In addition, it may even forge incorrect aggregated results to save resources. To deal with these issues, despite that the existing works can protect local model privacy and achieve verifiability of aggregated results, they are defective in protecting global model privacy, guaranteeing verifiability if collusion attacks occur, and suffer from high computation cost. To further tackle the above challenges, a verifiable and collusion-resistant secure aggregation scheme for FL is proposed, named VCSA. Concretely, we combine symmetric homomorphic encryption with single masking to protect model privacy. Meanwhile, we adopt verifiable multi-secret sharing and generalized Pedersen commitment to achieve verifiability and prevent users from uploading incorrect shares. Furthermore, high model accuracy can be ensured even if some users go offline. Security analysis illustrates that our VCSA enhances the security of FL, realizes verifiability despite collusion attacks and robustness to dropout. Performance evaluation displays that our VCSA can reduce at least 28.27% and 79.15% regarding computation cost compared to existing schemes.

An Efficient and Multi-Private Key Secure Aggregation Scheme for Federated Learning

An Efficient and Multi-Private Key Secure Aggregation Scheme for Federated Learning

CESA: Communication efficient secure aggregation scheme via sparse graph in federated learning

As a distributed learning paradigm, federated learning can be effectively applied to the decentralized system since it can resolve the “data island” problem. However, it is also vulnerable to serious privacy breaches. Although existing secure aggregation technique can address privacy concerns, they also incur significant additional computation and communication costs. To address these challenges, this paper offers a Communication Efficient Secure Aggregation scheme. Firstly, the central server uses the communication delay between terminals as the weight of the fully terminal-connected graph to transform it into a sparse connected graph based on the minimal spanning tree. Secondly, instead of relying on central server for key advertisement, the terminals advertise keys via a neighboring terminal forwarding approach based on sparsely graph. Thirdly, we propose using the central server for auxiliary advertising to address unexpected terminal dropout. Simultaneously, we theoretically demonstrate our scheme’s security and have lower computation and communication costs. Experiments show that CESA can reduce the running time by 28.2% without sacrificing security and model accuracy compared to conventional secure aggregation when there are 10 terminals in the system.

An efficient privacy-preserving and verifiable scheme for federated learning

As one of the most important methods of privacy computing, federated learning has attracted much attention as it makes data available but invisible (i.e., uploading gradients instead of raw data). However, adversaries may still recover some private information such as tabs, memberships or even training data, from gradients. Additionally, the malicious server may return the incorrect or forged aggregated result to clients for certain illegal interests. To ensure verifiability and privacy-preservation, in this paper, we present a verifiable secure aggregation scheme under the dual-server federated learning framework. Specifically, we combine the learning with error (LWE) cryptosystem with the secret sharing technique to guarantee the privacy of the aggregated result and each client’s local gradient. Meanwhile, we skillfully design a double-verification protocol, including the server-side and client-side verification, to efficiently verify the correctness of the aggregated result and ensure data availability. Specifically, two servers mutually verify the correctness of the aggregated result through the linear homomorphic hash technique. After passing the server-side mutual verification, the malicious server may still directly broadcast the forged aggregated result to clients. Our client-side verification protocol can ensure data availability to identify the correct aggregation result sent by the semi-trusted server. To the best of our knowledge, existing solutions do not take data availability into account. Extensive experimental comparisons with the state-of-the-art schemes demonstrate the effectiveness and efficiency of the proposed scheme in terms of accuracy, computational cost and communication overhead.

Smart contract assisted secure aggregation scheme for model update in federated learning

Federated learning (FL) is a cutting-edge machine learning architecture that effectively meets the data and model training requirements of numerous businesses, while ensuring privacy protection, data security, and adherence to legal constraints. However, recent studies have demonstrated that attackers can deduce users’ personal data from their shared model parameters, thus underscoring a substantial security risk for federated learning. Despite considerable efforts by academics to address this issue, current approaches suffer from drawbacks such as high communication and processing costs, insufficient robustness, and heavy reliance on trusted third parties. In this paper, we present the smart contract assisted secure aggregation scheme (SCSA) to overcome the aforementioned problems. First, we creatively offer a three-layer secure aggregation architecture that is particularly suitable for scenarios involving multiple sensors used in model training. Secondly, to ensure the credibility of the identities between participating entities in the model parameters aggregation process of FL, a novel certificateless authentication mechanism relying on certificateless encryption is provided. Moreover, for the sake of enhancing the security of the aggregation process in FL, we employ a combination of smart contracts and secret sharing techniques to distribute security masks to users in a decentralized and highly secure manner. It also combines with secure communication to create a double fault tolerance mechanism that significantly boosts the whole system’s robustness. The effectiveness of our approach in achieving secure aggregation is substantiated through theoretical analysis and simulated trials, while simultaneously ensuring strong security and robustness.

Federated Medical Learning Framework Based on Blockchain and Homomorphic Encryption

Federated learning-based medical data privacy sharing can promote the development of medical industry intelligence, but limited by its own security and privacy deficiencies, federated learning still suffers from a single point of failure and privacy leakage of intermediate parameters. To address these problems, this paper proposes a privacy protection framework for medical data based on blockchain and cross-silo federated learning, using cross-silo federated learning to establish a collaborative training platform for multiple medical institutions to enhance the privacy of medical data, introducing blockchain and smart contracts to realize decentralized federated learning to enhance trust between distrustful medical institutions and solve the problem of a single point of failure. In addition, a secure aggregation scheme is designed using threshold homomorphic encryption to prevent the privacy leakage problem during parameter transmission. The experimental and analytical results show that the accuracy of this paper’s scheme is consistent with the original federated learning scheme, effectively deals with the problems of single-point failure and inference attacks of federated learning, improves system robustness, and is suitable for medical scenarios with more stringent requirements on security and accuracy.

Lightweight and Verifiable Secure Aggregation for Multi-dimensional Data in Edge-enhanced IoT

The development of IoT applications promotes the generation of large-scale data. In this situation, more and more organizations or institutions are interested in collecting and analyzing these data to obtain high-quality information to serve people’s lives. Nevertheless, existing secure aggregation solutions are mostly proposed under the honest-but-curious threat model, without considering malicious aggregators. Actually, the aggregation node may return tampered or forged aggregation results for vulnerabilities or ulterior motives. It is more challenging to devise efficient verifiable data aggregation due to the characteristics of IoT devices. In this paper, we propose a Lightweight and Verifiable Secure Aggregation Scheme for Multi-dimensional Data in Edge-enhanced IoT (LVSA-MD), which allows the secure aggregation of multi-dimensional data, and the requester can verify the correctness and integrity of the aggregation results. In LVSA-MD, we combine Chinese remainder theorem with secret sharing technique to protect the individual data privacy, as well as reducing the communication overhead. Moreover, by involving homomorphic MAC, we achieve a lightweight verifiable mechanism to ensure the integrity of result. We additionally employ identity-based signature to achieve the source authentication. Detailed security analysis demonstrates that the proposed LVSA-MD achieves the goal of protecting privacy, integrity and authentication. Moreover, extensive theoretical analyses and experimental evaluations show that our LVSA-MD performs efficiently with respect to computation and communication while retaining more desired properties.

An Improved Federated Learning-Assisted Data Aggregation Scheme for Smart Grids

In the context of rapid advancements in artificial intelligence (AI) technology, new technologies, such as federated learning and edge computing, have been widely applied in the power Internet of Things (PIoT). When compared to the traditional centralized training approach, conventional federated learning (FL) significantly enhances privacy protection. Nonetheless, the approach poses privacy concerns, such as inferring other users’ training data through the global model or user-transferred parameters. In light of these challenges, this research paper introduces a novel privacy-preserving data aggregation scheme for the smart grid, bolstered by an improved FL technique. The secure multi-party computation (SMC) and differential privacy (DP) are skillfully combined with FL to combat inference attacks during both the learning process and output inference stages, thus furnishing robust privacy assurances. Through this approach, a trusted third party can securely acquire model parameters from power data holders and securely update the global model in an aggregated way. Moreover, the proposed secure aggregation scheme, as demonstrated through security analysis, achieves secure and reliable data aggregation in the electric PIoT environment. Finally, the experimental analysis shows that the proposed scheme effectively performs federated learning tasks, achieving good model accuracy and shorter execution times.

Olive: Oblivious Federated Learning on Trusted Execution Environment against the Risk of Sparsification

Combining Federated Learning (FL) with a Trusted Execution Environment (TEE) is a promising approach for realizing privacy-preserving FL, which has garnered significant academic attention in recent years. Implementing the TEE on the server side enables each round of FL to proceed without exposing the client's gradient information to untrusted servers. This addresses usability gaps in existing secure aggregation schemes as well as utility gaps in differentially private FL. However, to address the issue using a TEE, the vulnerabilities of server-side TEEs need to be considered---this has not been sufficiently investigated in the context of FL. The main technical contribution of this study is the analysis of the vulnerabilities of TEE in FL and the defense. First, we theoretically analyze the leakage of memory access patterns, revealing the risk of sparsified gradients, which are commonly used in FL to enhance communication efficiency and model accuracy. Second, we devise an inference attack to link memory access patterns to sensitive information in the training dataset. Finally, we propose an oblivious yet efficient aggregation algorithm to prevent memory access pattern leakage. Our experiments on real-world data demonstrate that the proposed method functions efficiently in practical scales.

CodedPaddedFL and CodedSecAgg: Straggler Mitigation and Secure Aggregation in Federated Learning

We present two novel federated learning (FL) schemes that mitigate the effect of straggling devices by introducing redundancy on the devices’ data across the network. Compared to other schemes in the literature, which deal with stragglers or device dropouts by ignoring their contribution, the proposed schemes do not suffer from the client drift problem. The first scheme, CodedPaddedFL, mitigates the effect of stragglers while retaining the privacy level of conventional FL. It combines one-time padding for user data privacy with gradient codes to yield straggler resiliency. The second scheme, CodedSecAgg, provides straggler resiliency and robustness against model inversion attacks and is based on Shamir’s secret sharing. We apply CodedPaddedFL and CodedSecAgg to a classification problem. For a scenario with 120 devices, CodedPaddedFL achieves a speed-up factor of 18 for an accuracy of 95% on the MNIST dataset compared to conventional FL. Furthermore, it yields similar performance in terms of latency compared to a recently proposed scheme by Prakash et al. without the shortcoming of additional leakage of private data. CodedSecAgg outperforms the state-of-the-art secure aggregation scheme LightSecAgg by a speed-up factor of 6.6–18.7 for the MNIST dataset for an accuracy of 95%.

Privacy-enhanced and non-interactive linear regression with dropout-resilience

Linear regression, the most basic and widely used machine learning algorithm, has played an essential role in many areas such as healthcare, economics, and weather prediction. However, in practice, regression training requires a large dataset, usually stored in a distributed form from multiple data owners. Moreover, it is difficult to federate data among users to train a global model because of privacy issues and communication limitations. To address these challenges, we propose a non-interactive privacy-enhanced training scheme for linear regression based on functional encryption - FELR. In particular, based on the secure aggregation scheme and Paillier inner-product functional encryption, two cloud servers can train the model on the ciphertext. The solution does not require any interaction between the cloud servers and data owners. Throughout the training period, the cloud servers do not know any local private data. A detailed security analysis demonstrates that FELR can provide high security for data owners. The experimental conclusions also demonstrate that the performance of the training scheme on the data owners' side is approximately 25% higher than that of the existing solution, which significantly decreases the computational costs.

Discrete GWO Optimized Data Aggregation for Reducing Transmission Rate in IoT

The conventional hospital environment is transformed into digital transformation that focuses on patient centric remote approach through advanced technologies. Early diagnosis of many diseases will improve the patient life. The cost of health care systems is reduced due to the use of advanced technologies such as Internet of Things (IoT), Wireless Sensor Networks (WSN), Embedded systems, Deep learning approaches and Optimization and aggregation methods. The data generated through these technologies will demand the bandwidth, data rate, latency of the network. In this proposed work, efficient discrete grey wolf optimization (DGWO) based data aggregation scheme using Elliptic curve Elgamal with Message Authentication code (ECEMAC) has been used to aggregate the parameters generated from the wearable sensor devices of the patient. The nodes that are far away from edge node will forward the data to its neighbor cluster head using DGWO. Aggregation scheme will reduce the number of transmissions over the network. The aggregated data are preprocessed at edge node to remove the noise for better diagnosis. Edge node will reduce the overhead of cloud server. The aggregated data are forward to cloud server for central storage and diagnosis. This proposed smart diagnosis will reduce the transmission cost through aggregation scheme which will reduce the energy of the system. Energy cost for proposed system for 300 nodes is 0.34μJ. Various energy cost of existing approaches such as secure privacy preserving data aggregation scheme (SPPDA), concealed data aggregation scheme for multiple application (CDAMA) and secure aggregation scheme (ASAS) are 1.3 μJ, 0.81 μJ and 0.51 μJ respectively. The optimization approaches and encryption method will ensure the data privacy.

Fog-enabled secure multiparty computation based aggregation scheme in smart grid

The Smart Grid (SG) is a modern power grid infrastructure for improved efficiency and reliability using modern telecommunications infrastructure, sensing, and measuring technologies. It provides sustainable electricity to customers and enables them to monitor their electricity use. The Fog Nodes (FNs) are deployed between the Cloud Control Center (CCC) and Smart Meters (SMs) to improve efficiency and reliability. The CCC is collecting usage data through FNs at regular intervals from SMs. As SM data contains private customer data, a secure scheme is required to ensure customer privacy during data collection. The usage view behavior of consumers is only possible if internal or external malicious entities collude together. In this paper, we present the Fog-Enabled Secure Multiparty Computation (FESMPC) based aggregation scheme for SG. This paper’s key contributions to the literature on the topic of secure privacy-aware aggregation schemes in SGs are threefold. Firstly, a secure aggregation scheme that preserves data privacy and ensures fault-tolerance is presented. The scheme is robust against the attacks of collusive and False Data Injection (FDI). A collusive attack is managed through enhanced Shamir’s secret scheme. At the same time, privacy during data aggregation is achieved through a Paillier homomorphic encryption scheme. Secondly, detailed security analyses are provided to show that the proposed scheme supports data privacy, source authentication, fault tolerance, and prevents FDI attacks. Thirdly, the paper presents extensive performance evaluations demonstrating that the proposed program scheme is efficient in encryption, aggregation, decryption, and less communication costs than the three existing state-of-the-art programs.

A Novel Secure Data Aggregation Scheme Based on Semi-Homomorphic Encryption in WSNs

Privacy protection in Wireless Sensor Networks (WSN) constitutes a big challenge for the adoption of WSNs in data sensitive applications like health monitoring or tracking and surveillance of borders. Privacy protection require additional controls and communications overloads, which impact the overall network lifetime. Research community has proposed several scenarios to minimize the impact of data protection generally based on secure aggregation and encryption to meet the practical requirements of energy constraints imposed by WSN. However, efficiency of privacy protection must be assessed before deployment. The privacy protection mechanisms are evaluated based on their hackability and network performance using four main metrics: Control Packet Overhead, delay, Throughput, Packet delivery ratio. The purpose of this paper is to propose a secure aggregation scheme based on homomorphic encryption. The new scheme will be will be compared to another scheme based on network metric and attack detection accuracy to have full view on the scheme performance for both network and security metrics. The proposed scheme named “Cluster-based Semi-Homomophic Encryption Aggregated Data” (CSHEAD) offer better performance as it reduces the controls overhead with higher detection accuracy. The conducted simulations confirm the expected results.

Multi-functional secure data aggregation schemes for WSNs

Secure data aggregation schemes are widely adopted in wireless sensor networks, not only to minimize the energy and bandwidth consumption, but also to enhance the security. Statistics obtained from data aggregation schemes often fall into three categories, i.e., distributive, algebraic, and holistic. In practice, a wide range of reasonable aggregation queries are combinations of several different statistics. Providing multi-functional aggregation support is also a primary demand for data preprocessing in data mining. However, most existing secure aggregation schemes only focus on a single type of statistics. Some statistics, especially holistic ones (e.g., median), are often difficult to compute efficiently in a distributed mode even without considering the security issue. In this paper, we first propose a new Multi-functiOnal secure Data Aggregation scheme (MODA), which encodes raw data into well-defined vectors to provide value-preservation, order-preservation and context-preservation, and thus offering the building blocks for multi-functional aggregation. A homomorphic encryption scheme is adopted to enable in-ciphertext aggregation and end-to-end security. Then, two enhanced and complementary schemes are proposed based on MODA, namely, RandOm selected encryption based Data Aggregation (RODA) and COmpression based Data Aggregation (CODA). RODA can significantly reduce the communication cost at the expense of slightly lower but acceptable security on a leaf node, while CODA can dramatically reduce communication cost with the lower aggregation accuracy. The performance results obtained from theoretic analysis and experimental evaluation of three real datasets under different scenarios, demonstrate that our schemes can achieve the performance superior to the most closely related work.

Two Is Not Enough: Privacy Assessment of Aggregation Schemes in Smart Metering

Abstract The widespread deployment of smart meters that frequently report energy consumption information, is a known threat to consumers’ privacy. Many promising privacy protection mechanisms based on secure aggregation schemes have been proposed. Even though these schemes are cryptographically secure, the energy provider has access to the plaintext aggregated power consumption. A privacy trade-off exists between the size of the aggregation scheme and the personal data that might be leaked, where smaller aggregation sizes leak more personal data. Recently, a UK industrial body has studied this privacy trade-off and identified that two smart meters forming an aggregate, are sufficient to achieve privacy. In this work, we challenge this study and investigate which aggregation sizes are sufficient to achieve privacy in the smart grid. Therefore, we propose a flexible, yet formal privacy metric using a cryptographic game based definition. Studying publicly-available, real world energy consumption datasets with various temporal resolutions, ranging from minutes to hourly intervals, we show that a typical household can be identified with very high probability. For example, we observe a 50% advantage over random guessing in identifying households for an aggregation size of 20 households with a 15-minutes reporting interval. Furthermore, our results indicate that single appliances can be identified with significant probability in aggregation sizes up to 10 households.

Anonymous and secure aggregation scheme in fog-based public cloud computing

By using fog computing, cloud computing can be extended to the edge of the network. Generally, in the public cloud, fog computing comprises three components: terminal device, fog node and public cloud server (PCS). In this paper, we propose the concept of anonymous and secure aggregation scheme (ASAS) in fog-based public cloud computing. In the ASAS model, a fog node aggregates the data from terminal nodes and forwards the aggregated data to the public cloud server. By using the ASAS scheme, the fog node can help terminal devices upload their data to PCS. By using the data aggregation technique, our ASAS scheme can save bandwidth between the fog node and PCS. At the same time, our ASAS scheme not only protects the identities of terminal devices by using pseudonyms but it also guarantees data secrecy via a homomorphic encryption technique. In this paper, we design the first concrete ASAS scheme. We also examine the security and the performance of our proposal, which we show to be provably secure and efficient.

A novel secure aggregation scheme for wireless sensor networks using stateful public key cryptography

Wireless sensor networks (WSNs) are nowadays considered as an important part of the Internet of Things (IoT). In these networks, data aggregation plays an essential role in energy preservation. However, WSNs are usually deployed in hostile and unattended environments (e.g. military applications) in which the confidentiality and integrity security services are widely desired. Recently, homomorphic encryptions have been applied to conceal sensitive information during aggregation such that algebraic operations are done directly on ciphertexts without decryption. The main benefit is that they offer the end-to-end data confidentiality and they do not require expensive computation at aggregator nodes since no encryption and decryption are performed. However, existing solutions either incur a considerable overhead or have limited applicability to certain types of aggregate queries. This paper presents a novel secure data aggregation protocol for WSNs. The scheme employs Stateful Public Key Encryption (StPKE) and some previous techniques in order to provide an efficient end-to-end security. Moreover, our solution does not impose any bound on the aggregation function’s nature (Maximum, Minimum, Average, etc.). We present and implement our scheme on TelosB as well as MicaZ sensor network platforms and measure the execution time of our various cryptographic functions. Simulations are also conducted to show how our scheme can achieve a high security level (by providing the above security services) with a low overhead (in terms of computation and communication) in large-scale scenario.

Secure Continuous Aggregation in Wireless Sensor Networks

Continuous aggregation is usually required in many sensor applications to obtain the temporal variation information of aggregates. However, in a hostile environment, the adversary could fabricate false temporal variation patterns of the aggregates by manipulating a series of aggregation results through compromised nodes. Existing secure aggregation schemes conduct one individual verification for each aggregation result, which could incur great accumulative communication cost and negative impact on transmission scheduling for continuous aggregation. In this paper, we identify distinct design issues for protecting continuous in-network aggregation and propose a novel scheme to detect false temporal variation patterns. Compared with the existing schemes, our scheme greatly reduces the verification cost by checking only a small part of aggregation results to verify the correctness of the temporal variation patterns in a time window. A sampling-based approach is used to check the aggregation results, which enables our scheme independent of any particular in-network aggregation protocols as opposed to existing schemes. We also propose a series of security mechanisms to protect the sampling process. Both theoretical analysis and simulations show the effectiveness and efficiency of our scheme.

ID List Forwarding Free Confidentiality Preserving Data Aggregation for Wireless Sensor Networks

Wireless sensor networks (WSNs) are composed of sensor nodes with limited energy which is difficult to replenish. Data aggregation is considered to help reduce communication overhead with in-network processing, thus minimizing energy consumption and maximizing network lifetime. Meanwhile, it comes with challenges for data confidentiality protection. Many existing confidentiality preserving aggregation protocols have to transfer list of sensors' ID for base station to explicitly tell which sensor nodes have actually provided measurement. However, forwarding a large number of node IDs brings overwhelming extra communication overhead. In this paper, we propose provably secure aggregation scheme perturbation-based efficient confidentiality preserving protocol (PEC2P) that allows efficient aggregation of perturbed data without transferring any ID information. In general, environmental data is confined to a certain range; hence, we utilize this feature and design an algorithm to help powerful base station retrieve the ID of reporting nodes. We analyze the accuracy of PEC2P and conclude that base station can retrieve the sum of environmental data with an overwhelming probability. We also prove that PEC2P is CPA secure by security reduction. Experiment results demonstrate that PEC2P significantly reduces node congestion (especially for the root node) during aggregation process in comparison with the existing protocols.