Data Protection: Combining Fragmentation, Encryption, and Dispersion, an Intermediary report

Gérard Memmi, Katarzyna Kapusta, Han QiuarXiv 2015 · OpenReview

TL;DR

This paper proposes a comprehensive data protection system (FEDS) that combines fragmentation, encryption, and dispersion to secure data throughout its lifecycle. By leveraging techniques like selective encryption and GPU acceleration, the authors aim to achieve robust security with improved performance and cost-effectiveness, addressing the limitations of traditional encryption-only approaches.

What Problem Does This Paper Address?

In today's digital landscape, data is constantly under threat from sophisticated attacks. Relying solely on encryption is insufficient due to its computational complexity and the ongoing advancements in cryptanalysis. The paper addresses the critical need for a more robust and cost-effective data protection strategy that can safeguard information across various stages of its lifecycle – from creation to storage and transmission. This includes protecting against data theft, alteration, and destruction, especially in the context of cloud computing and distributed systems where data is increasingly vulnerable.

Methods

The core of the proposed solution is a Fragmentation, Encryption, and Dispersion System (FEDS). Key methods include:

  • Fragmentation: Data is broken down into smaller pieces (fragments). The paper explores two main types:
  • Selective Fragmentation: Information is split into a public and a private fragment, allowing for differentiated protection levels.
  • Non-Selective Fragmentation: Data is treated as a sequence of bits without structural consideration.
  • Encryption: Fragments are encrypted, with the strength of encryption tailored to the fragment's criticality. The paper highlights selective encryption, particularly for images, where only critical frequency domain coefficients are encrypted.
  • Dispersion: Fragments are distributed across multiple physical machines or cloud environments.
  • GPU Acceleration: The authors extensively leverage General Purpose Graphics Processing Units (GPUs) to accelerate computationally intensive tasks like Discrete Cosine Transform (DCT) and Discrete Wavelet Transform (DWT), significantly improving performance.
  • DWT for Lossless Processing: The paper introduces the use of Discrete Wavelet Transform (DWT) with a Le Gall 5/3 filter for image processing, emphasizing its lossless property to avoid rounding errors and ensure data integrity, a limitation of DCT.

Datasets

  • Bitmap Images: Used for demonstrating selective encryption techniques. Specific images mentioned include "Barbara" and "baboon."
  • La Poste Data: Provided by the ITEA2 CAP project, used for scalability experiments.
  • Text Files: Used for security analysis alongside image data.

No newly introduced benchmarks are explicitly mentioned, but the paper compares its performance against standard AES implementations.

Key Insights and Ideas

  • Beyond Encryption: The paper strongly advocates for a multi-layered approach (fragmentation, encryption, dispersion) as a more effective and cost-efficient solution than relying solely on encryption.
  • Selective Protection is Key: Differentiating data based on confidentiality and applying protection accordingly is crucial for optimizing performance and resource usage.
  • GPU Power for Security: GPUs are not just for graphics; they offer significant parallel processing capabilities that can dramatically speed up cryptographic operations, making advanced security techniques more feasible.
  • DWT for Integrity: The lossless nature of DWT is a critical advantage for applications requiring high data integrity, overcoming limitations of DCT-based methods.
  • Agnostic Design: The proposed methods are designed to be agnostic to data type, making them broadly applicable across various systems and data formats.

Key Results

  • Performance Gains: Selective encryption using GPU acceleration achieved significant speedups, with DCT 8x8 being up to 70 times faster on GPU than on CPU. The DWT-based method also showed substantial acceleration.
  • Security Robustness: Statistical analysis (histogram uniformity, correlation coefficients, entropy) demonstrated that the proposed scheme effectively removes statistical patterns and spatial redundancy, making encrypted data appear random and resistant to common cryptanalytic attacks.
  • Data Integrity: The use of DWT with the Le Gall 5/3 filter ensured lossless processing for images, avoiding rounding errors and maintaining data integrity.
  • Efficiency: The selective approach significantly reduced the amount of data that needed to be encrypted, leading to improved performance and reduced storage overhead. For example, the confidential fragment in image processing was typically less than 10% of the total data.

Limitations and Future Work

  • Performance Bottlenecks: While GPU acceleration is effective, memory transfer between GPU and host memory can still be a bottleneck.
  • Key Management: The paper acknowledges the ongoing challenge of key management, especially for long-term storage, and suggests further research into keyless solutions or more robust key management strategies.
  • Formal Security Proofs: The authors suggest the need for formal methods to compare different FEDS approaches and set optimal parameters for desired security levels.
  • Broader Data Types: While the proposed methods are designed to be agnostic, further experiments with diverse data formats (HD images, audio, video) are recommended.
  • Transmission Security: The current work focuses on data storage; future work will explore using fragmentation for data transmission security.
  • Defragmentation Avoidance: Investigating techniques like multiparty computation or homomorphic encryption to process data without full defragmentation is a promising avenue.

Summary updated

Abstract

Hardening data protection using multiple methods rather than 'just' encryption is of paramount importance when considering continuous and powerful attacks in order to observe, steal, alter, or even destroy private and confidential information.Our purpose is to look at cost effective data protection by way of combining fragmentation, encryption, and dispersion over several physical machines. This involves deriving general schemes to protect data everywhere throughout a network of machines where they are being processed, transmitted, and stored during their entire life cycle. This is being enabled by a number of parallel and distributed architectures using various set of cores or machines ranging from General Purpose GPUs to multiple clouds. In this report, we first present a general and conceptual description of what should be a fragmentation, encryption, and dispersion system (FEDS) including a number of high level requirements such systems ought to meet. Then, we focus on two kind of fragmentation. First, a selective separation of information in two fragments a public one and a private one. We describe a family of processes and address not only the question of performance but also the questions of memory occupation, integrity or quality of the restitution of the information, and of course we conclude with an analysis of the level of security provided by our algorithms. Then, we analyze works first on general dispersion systems in a bit wise manner without data structure consideration; second on fragmentation of information considering data defined along an object oriented data structure or along a record structure to be stored in a relational database.
BibTeX
@misc{memmi2015data,
  author = {Gérard Memmi and Katarzyna Kapusta and Han Qiu},
  title = {{Data Protection: Combining Fragmentation, Encryption, and Dispersion, an Intermediary report}},
  year = {2015},
  eprint = {1512.02951},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/1512.02951},
}
Figure