This thesis presents a comprehensive analysis of Fully Homomorphic Encryption (FHE), a transformative advancement in contemporary cryptography. The exponential growth in digital data, coupled with the escalating demand for remote data processing, has elevated the importance of robust information security and privacy mechanisms. Homomorphic Encryption (HE) schemes facilitate the outsourcing of computations to untrusted third parties without exposing the underlying sensitive information, significantly enhancing data privacy assurances. Traditional encryption methods effectively protect data at rest and in transit; however, they inherently lack the capability to perform operations directly on encrypted data. FHE addresses this fundamental shortcoming by enabling intricate computations to be conducted on encrypted data without the necessity of decryption. The thesis begins by outlining the historical development of encryption systems and fundamental cryptographic concepts. It details the progression from the "privacy homomorphism" concept introduced by Rivest, Adleman, and Dertouzos in 1978 to Craig Gentry's seminal work in 2009, which presented the first practical FHE scheme. The key milestones in the evolution of FHE and major contributions in the literature are presented from both theoretical and practical perspectives. Subsequently, the mathematical structures and concepts underlying FHE are discussed in depth. Topics such as lattices, rings and fields, polynomials, number fields, canonical embedding, and ideal structures are examined as the foundational building blocks ensuring the security of FHE. In particular, challenging problems provided by Lattice-Based Cryptography, such as Learning with Errors (LWE), serve as the mathematical basis for FHE's resistance to attacks. The thesis also analyzes the structure, functioning, and differences between modern FHE schemes (such as Brakerski, Gentry and Vaikuntanathan (BGV), Fan and Vercauteren (BFV), and Cheon-Kim-Kim-Song (CKKS)). The practical applications, advantages, limitations, and recent optimization techniques of these schemes are thoroughly examined. Special attention is given to the opportunities FHE offers in sectors where data privacy is paramount, including healthcare, finance, and cloud computing. Key practical challenges such as high computational costs, large key and ciphertext sizes, and noise management are highlighted. Innovations to address these challenges, including hardware accelerators, software optimizations, and open-source libraries (like HELib and SEAL), are explored in detail. The thesis also covers contemporary research areas such as multi-key FHE, privacy-preserving computation in artificial intelligence, and post-quantum cryptography. In conclusion, this thesis presents a broad perspective on FHE, covering its mathematical foundations and practical application potential, while evaluating current developments, limitations, and future research directions. The wider and more efficient xxv adoption of FHE will open new possibilities for information security and privacy. In this context, the thesis aims to serve as a comprehensive resource for both academic and industrial communities. Keywords: Bootstrapping, Relinearization, Ring Learning with Errors, BGV, BFV, CKKS.
This thesis presents a comprehensive analysis of Fully Homomorphic Encryption (FHE), a transformative advancement in contemporary cryptography. The exponential growth in digital data, coupled with the escalating demand for remote data processing, has elevated the importance of robust information security and privacy mechanisms. Homomorphic Encryption (HE) schemes facilitate the outsourcing of computations to untrusted third parties without exposing the underlying sensitive information, significantly enhancing data privacy assurances. Traditional encryption methods effectively protect data at rest and in transit; however, they inherently lack the capability to perform operations directly on encrypted data. FHE addresses this fundamental shortcoming by enabling intricate computations to be conducted on encrypted data without the necessity of decryption. The thesis begins by outlining the historical development of encryption systems and fundamental cryptographic concepts. It details the progression from the "privacy homomorphism" concept introduced by Rivest, Adleman, and Dertouzos in 1978 to Craig Gentry's seminal work in 2009, which presented the first practical FHE scheme. The key milestones in the evolution of FHE and major contributions in the literature are presented from both theoretical and practical perspectives. Subsequently, the mathematical structures and concepts underlying FHE are discussed in depth. Topics such as lattices, rings and fields, polynomials, number fields, canonical embedding, and ideal structures are examined as the foundational building blocks ensuring the security of FHE. In particular, challenging problems provided by Lattice-Based Cryptography, such as Learning with Errors (LWE), serve as the mathematical basis for FHE's resistance to attacks. The thesis also analyzes the structure, functioning, and differences between modern FHE schemes (such as Brakerski, Gentry and Vaikuntanathan (BGV), Fan and Vercauteren (BFV), and Cheon-Kim-Kim-Song (CKKS)). The practical applications, advantages, limitations, and recent optimization techniques of these schemes are thoroughly examined. Special attention is given to the opportunities FHE offers in sectors where data privacy is paramount, including healthcare, finance, and cloud computing. Key practical challenges such as high computational costs, large key and ciphertext sizes, and noise management are highlighted. Innovations to address these challenges, including hardware accelerators, software optimizations, and open-source libraries (like HELib and SEAL), are explored in detail. The thesis also covers contemporary research areas such as multi-key FHE, privacy-preserving computation in artificial intelligence, and post-quantum cryptography. In conclusion, this thesis presents a broad perspective on FHE, covering its mathematical foundations and practical application potential, while evaluating current developments, limitations, and future research directions. The wider and more efficient xxv adoption of FHE will open new possibilities for information security and privacy. In this context, the thesis aims to serve as a comprehensive resource for both academic and industrial communities. Keywords: Bootstrapping, Relinearization, Ring Learning with Errors, BGV, BFV, CKKS.
Fully homomorphic encryption (FHE) allows arbitrary computations to be performed directly on encrypted data and is a core technology for privacy computing and secure cloud computing. The Number Theory Research Unit (NTRU) has become a research topic in this field by virtue of its lattice-based anti-quantum security, efficient polynomial operation characteristics and natural multi-key homomorphism capability. In this paper, we systematically review the key advances of NTRU-based FHE, covering the directions of the design of FHE, the optimization of bootstrapping techniques, and the application to FHE for multi-key approaches. Furthermore, the innovativeness and practical application challenges of the schemes are analyzed, and the future research trends are envisioned.
Since Gentry's breakthrough construction of fully homomorphic encryption from lattice-based assumptions (STOC 2009), homomorphic cryptography has attracted a lot of attention. In short, homomorphic cryptography schemes allow performing computation on encrypted data without knowing anything about the underlying plaintext. This branch of cryptography has become increasingly useful in building new protocols and schemes with intriguing security and functionality features. In this thesis, we continue to study the applications of homomorphic cryptography and the lattice-based techniques underlying them in realizing new and enhanced cryptographic primitives. We obtain the following results: -We construct the first noninteractive zero knowledge argument (and proof) system for all of NP from standard lattice assumptions. Noninteractive zero knowledge argument systems have found many applications in enhancing the functionality as well as the security of cryptographic schemes and protocols. Constructing noninteractive zero knowledge arguments from lattice assumption has been a long standing open question. We finally close this problem. -We consider multi-key fully homomorphic encryption (FHE) schemes. Traditional fully homomorphic encryption schemes allow computation on plaintext encrypted under a single key. The notion of multi-key fully homomorphic encryption allows homomorphic computation on data encrypted under different keys. We construct multi-key FHE schemes which are naturally dynamic: ciphertexts under new keys can join even during the homomorphic computation. -Finally, we focus on constrained pseudorandom functions (C-PRFs), which are pseudorandom functions (PRFs) with additional functional capabilities. We propose a new approach for building C-PRFs from lattices, and also significantly enhance the underlying lattice parameters.
Abstract The rapid evolution of quantum computing fundamentally threatens modern cryptographic infrastructures, necessitating a paradigm shift toward advanced encryption techniques. While post-quantum cryptography (PQC) ensures that data remains secure against quantum algorithmic decryption, it does not inherently facilitate the processing of that data in untrusted environments. Homomorphic encryption (HE) addresses this computational privacy requirement, yet traditional HE schemes rely on classical mathematical problems that are easily compromised by quantum systems. This paper presents a comprehensive comparative framework that integrates post-quantum cryptographic principles with homomorphic capabilities to enable quantum-resistant, privacy-preserving computation. By analyzing lattice-based algorithms, code-based cryptography, and hardware acceleration paradigms, this study maps the existing trade-offs between security, computational overhead, and practical deployment feasibility in the impending quantum era. Keywords: PQC, HE. 1.Introduction The rapid evolution of quantum computing technology presents a profound threat to modern cryptographic infrastructures. As quantum algorithms, notably Shor's algorithm, mature, they acquire the capability to solve nondeterministic polynomial time problems in polynomial time, thereby breaking traditional asymmetric encryption standards such as RSA and Elliptic Curve Cryptography (ECC) (Chen, 2024). Consequently, the development of post-quantum cryptography (PQC) has become a global imperative to secure sensitive communications and data against future quantum adversaries (Pranjal & Chaturvedi, 2024). Cryptographers are currently focusing on transitioning secure systems to these newly established PQC standards before large-scale quantum computers become commercially viable. While PQC addresses the secure transmission and storage of data, it does not inherently solve the problem of secure data processing in untrusted cloud environments. Homomorphic encryption (HE) provides a sophisticated mechanism that allows computations to be performed directly on encrypted data without requiring prior decryption (Jain & Cherukuri, 2023). The core problem addressed in this paper is the conceptual and practical integration of PQC with HE shown in Fig1.1 which is essential for ensuring quantum-safe, privacy-preserving computation in distributed architectures. Achieving this synthesis is highly complex due to the massive computational overhead and noise accumulation inherent to both cryptographic domains. However, existing approaches to privacy-preserving computation remain broadly insufficient for the post-quantum era. First, conventional homomorphic encryption schemes rely heavily on mathematical hardness assumptions, such as integer factorization or discrete logarithms, which are demonstrably vulnerable to quantum attacks (Chen, 2024). Second, while some high-performance fully homomorphic implementations exist, they suffer from prohibitive computational overhead and severe ciphertext expansion, making widespread deployment impractical without specialized mitigation strategies (Tseng et al., 2025). Third, hardware-accelerated privacy solutions often rely on trusted execution environments, like Intel SGX, which inherently introduce side-channel vulnerabilities and force reliance on centralized hardware trust rather than pure mathematical security (Sadat et al., 2017). To overcome these critical limitations, this paper proposes a structured comparative methodology for evaluating quantum-resistant homomorphic encryption systems. Specifically, our paper makes the following main contributions: We provide a comprehensive comparative framework that categorizes and evaluates lattice-based and code-based homomorphic encryption schemes against classical performance constraints. We propose a structured, hypothetical evaluation pipeline designed to benchmark the computational overhead and latency of post-quantum homomorphic algorithms in federated networking environments. 2.Privacy-preserving computation The literature surrounding privacy-preserving computation can be broadly classified into three primary categories, the first being classical and fully homomorphic encryption (FHE) systems. The core idea behind these schemes is the preservation of algebraic structures during encryption, allowing operations on ciphertexts to reflect directly on the underlying plaintexts without exposing the raw data (Jain & Cherukuri, 2023). A major strength of these classical FHE systems, such as those utilized in encrypted relational databases like FHE-SQL, is their ability to provide end-to-end cryptographic protection for queries without relying on trusted hardware modules (Tseng et al., 2025). However, a fundamental weakness is that early FHE models, and their underlying privacy homomorphisms, remain fundamentally susceptible to quantum cryptanalysis due to their reliance on classical mathematical problems (Sen, 2013). Compared to these classical paradigms, our work specifically focuses on migrating these operational homomorphic properties into genuinely quantum-safe mathematical frameworks. The second category of related work focuses on hardware acceleration and hybrid computational environments for homomorphic encryption. The core idea here involves leveraging specialized hardware, such as ASIC AI chips or coarse-grained compute architectures like Google's TPUs, to mitigate the enormous computational overhead associated with HE (Tong et al., 2025). The strength of this approach lies in its ability to bridge the energy-efficiency gap, often utilizing intelligent compiler frameworks to convert high-precision modular arithmetic into dense, low-precision matrix multiplications. Conversely, the weakness of hardware-centric methods is that architectural mismatches frequently lead to severe resource under-utilization, and hybrid solutions utilizing components like Intel SGX introduce hardware trust assumptions that can compromise pure cryptographic integrity (Sadat et al., 2017). Unlike these approaches, our framework prioritizes algorithmic resilience and quantum resistance over purely hardware-driven and enclave-based optimizations. The third category encompasses emerging post-quantum and quantum homomorphic paradigms, which seek to unify quantum resistance with computational privacy. Researchers have proposed lattice-based HE schemes to replace vulnerable legacy systems like Paillier and ElGamal, successfully enabling secure applications such as federated learning (Chen, 2024). Similarly, code-based cryptography utilizing McEliece and non-negative matrix factorization has shown promise in providing faster encryption and decryption times while remaining mathematically secure against quantum computers (Chen, 2024). Furthermore, advanced quantum homomorphic encryption protocols have explored information-theoretic security combined with homomorphic quantum error-correction to ensure future-proof data protection (Ouyang & Rohde, 2022). While these systems represent the state-of-the-art, their primary weakness remains the massive memory and communication overhead introduced during distributed training protocols, which negatively impacts classification performance. Our approach builds upon this category by proposing a standardized methodology to systematically compare these disparate post-quantum homomorphic techniques. 3.Multi-stage Method To systematically evaluate and deploy post-quantum homomorphic encryption, we propose a structured, multi-stage methodological framework. The first step involves cryptographic primitive selection, where system architects must evaluate and choose between lattice-based constructs, such as ML-KEM adaptations, and code-based systems like McEliece, depending on their domain constraints (Chen, 2024). The second step requires the implementation of an algorithmic parameter optimization module, which dynamically adjusts polynomial degrees and error distributions to balance security levels with the operational multiplicative depth required by the application. Finally, the third step comprises the integration of a specialized abstraction layer designed to map complex high-precision cryptographic operations into hardware-friendly matrix operations, thereby reducing the inherent bottleneck of ciphertext evaluation. The key design choices within this framework are driven by the absolute necessity to balance quantum resistance with practical algorithmic utility. We prioritize lattice-based cryptography for our primary evaluation pathway because of its established resilience against quantum threats and its recent standardization by the National Institute of Standards and Technology (NIST) (Chen, 2024). Additionally, we explicitly design our framework to avoid reliance on trusted hardware environments, opting instead for pure cryptographic guarantees to prevent the side-channel data leaks that often plague hardware-enclave solutions. By focusing on algorithmic alignment and basis-aligned mathematical transformations, the framework ensures that homomorphic workloads can be optimized effectively without compromising the underlying post-quantum security proofs. To validate the efficacy of the proposed framework, we outline a comprehensive evaluation plan utilizing a hypothetical, high-dimensional healthcare benchmarking dataset. The experimental setup will simulate a secure genome-wide association study (GWAS) executing within a fully decentralized federated learning environment (Sadat et al., 2017). The primary metrics for this hypothetical evaluation will include key generation latency, ciphertext expansion ratio, and the amortized computational time required for homomorphic multiplication operations across distributed nodes. By systematically varying the cryptographic noise parameters and observing the resulting degradation in analytical performance, we aim to empirically map the trad
The idea of computing on encrypted data without decryption dates back to the notion of privacy homomorphisms introduced by Rivest, Adleman, and Dertouzos (1978). Their proposals built using the elegant structure of Chinese Remainder Theorem (CRT), were later shown to be insecure under simple known-plaintext attacks. Subsequent CRT-based fully homomorphic encryption (FHE) over the integers addresses this algebraic transparency by injecting noise and basing security on approximate common divisor–type assumptions, but the resulting designs are burdened by large public keys and costly ciphertext refresh procedures. In this work, we develop a new CRT-based FHE scheme whose security relies on the Ring-LWE (RLWE) hardness assumption. For this purpose, we introduce the CRT-RLWE problem. We show that the problem is at least as hard as the RLWE, thereby positioning our construction within the established post-quantum security landscape of RLWE-based cryptography. Our scheme retains an explicit CRT embedding, separating a message component modulo a prime-power plaintext modulus and an auxiliary CRT component, while using RLWE-style key and ring arithmetic for compactness and efficiency. Finally, we make it fully homomorphic by using a new bootstrapping procedure, that adopts the recryption paradigm for BGV/BFV schemes utilizing the linear transformation and digit extraction techniques.
Web-servers are a dominant medium of modern communication and thus process vast amounts of highly sensitive, private data. Social networks, online auctions, and all kinds of cloud services constitute popular examples of online services complying with this paradigm. Due to many alarming scandals regarding the misuse of private data, IT security solutions protecting this sensitive data (for example by hiding it through encryption) are indispensable. Ideally, web-servers should be able to compute on private data without seeing it - we refer to this as the "Secure Outsourcing of Computation". Most notably, the cryptographic primitive of Homomorphic Encryption (HE) enables the web-servers to process data in the encrypted domain, which thereby hides/protects all private data as it is never available in the clear. More precisely, HE is a type of encryption that allows for the computation of certain functionalities on encrypted data without knowledge of the decryption key. Depending on the HE scheme, various functionalities can be evaluated in the encrypted domain, ranging from additions or multiplications only (Group Homomorphic Encryption, GHE), to virtually any computable functionality (Fully Homomorphic Encryption, FHE). We provide a clean foundational framework for HE that permits security characterizations for a large class of HE schemes and shows strong similarities between GHE and FHE. We study these characterizations in order to assess the limits of HE, and show that GHE is obsolete in the quantum world, meaning that any GHE scheme can be broken by a quantum computer. Together with our newly constructed universal blueprint that encompasses virtually any GHE scheme, we know exactly how these schemes must be designed and what the underlying hardness assumptions must look like. To some extent, this rounds off the topic of GHE both in the standard and quantum model of computation. In the standard model (which is presently assumed to be the most realistic model of computation), GHE is still highly attractive as it provides very efficient solutions to a variety of practical problems. On the basis of our foundational framework, we construct new GHE schemes with unique features and show their employability in practical applications. Most importantly, we use the special features of one of these schemes and develop a novel technique to achieve a practically efficient solution to the problem of securely outsourcing computations. Our construction allows multiple users to send encryptions of their private data under their own keys to a distrusted web-server that can then, on behalf of the users, evaluate any dynamically chosen computable function on these inputs in the encrypted domain (even across encryptions under different public keys) without learning anything at all, and without interacting with the users.
Homomorphic encryption is a recent and fundamental breakthrough in modern cryptography, which allows the performance of operations on encrypted data without unveiling the data. Leveraging quantum mechanics principles, quantum computers can potentially solve certain computational problems exponentially faster than classical computers. This immense computational power offers new possibilities for various fields, including cryptography. The rapid evolution of both these fields has led to the development of quantum fully homomorphic encryption (QFHE), which makes the capabilities of classical HE extend into the quantum domain. However, many existing QFHE schemes require significant memory due to complex calculations and fault-tolerance needs. This paper contributes in two ways. First, we provide a comprehensive survey of two specific QFHE schemes, discussing their underlying principles, mathematical frameworks, security aspects, and practical applications. We also explore the challenges posed by quantum computing and how QFHE addresses these to achieve both security and computational efficiency. Second, we propose a new hierarchical memory management system for QFHE, which includes a “quantum cache” (a specialized memory storage for quantum data) and a “reinforcement learning agent” (an intelligent system that learns from experience to optimize decisions). This system dynamically manages data movement between the cache and classical memory, improving memory efficiency and potentially boosting computational performance.