A Heterogeneous Database Encryption Algorithm Based on B/S Structure

Jun Chen, Feng ZhangProceedings of the International Conference on Information Economy, Data Modeling and Cloud Computing, ICIDC 2022, 17-19 June 2022, Qingdao, China 2022 · DOI

This paper, presented at the ICIDC 2022 conference, proposes a new way to secure heterogeneous databases—collections of multiple, associated databases that allow for shared data access. As these systems become more common in cloud environments, protecting them without sacrificing performance has become a significant hurdle.

TL;DR

The authors introduce a heterogeneous database encryption algorithm built on a Browser/Server (B/S) architecture. By implementing a multi-level key management system and access policy trees, the proposed method eliminates data loss during encryption and significantly reduces response times for data queries and updates compared to traditional methods.

What Problem Does This Paper Address?

Heterogeneous databases are powerful because they allow different, autonomous database systems to talk to each other. However, this complexity makes them prime targets for data breaches.

The authors identify a specific bottleneck in traditional encryption: "uneven key correspondence" between the user and the central authority. In older Client/Server (C/S) models, this often leads to long encryption times and high latency. Furthermore, some existing encryption algorithms (like those based on "association operations") can actually cause data loss during the transformation process. The challenge is to create a system that is secure, maintains data integrity, and remains fast enough for real-world use.

Methods

The researchers shifted the architecture from the traditional C/S model to a Browser/Server (B/S) structure. In this setup, the primary transaction logic and encryption processes happen on the server side, which simplifies the client-side requirements and lowers maintenance costs.

The core of their approach involves two main components:

  1. Hierarchical Key Management: Instead of a flat structure, they use a multi-level system. It starts with a Central Authority ($TA$), which manages various Permission Centers ($AA$). These centers then generate specific keys for users. The paper uses bilinear groups and mappings to mathematically define the system's public and master keys.
  2. Access Policy Tree Encryption: To handle complex permissions, the algorithm uses a policy tree. Plaintext data is converted into ciphertext based on a random number ($\theta$) and a specific access strategy. They utilize Lagrange polynomials to determine threshold polynomials, ensuring that only users with the correct attributes and keys can reconstruct the data.

Datasets

The paper evaluates the algorithm through a simulation. While it does not name a specific public dataset (such as a standard benchmark like TPC-H), it uses a representative "heterogeneous database" to compare the proposed algorithm against two others:

  • MA: A database encryption algorithm based on association operations.
  • MB: A database encryption algorithm based on the Arnold transform.

Key Insights and Ideas

  • Zero Data Loss: One of the most critical findings is that the proposed B/S algorithm ensures 100% data integrity. Unlike the "MA" algorithm, which showed significant data loss in visual simulations (Figure 2), the proposed method preserves all original information.
  • Server-Side Heavy Lifting: By moving the encryption logic to the WEB server within the B/S structure, the system reduces the "unevenness" of key management that plagues traditional methods.
  • Scalability via Hierarchy: The use of recursive depth in the key structure allows the system to scale its permission levels without a linear increase in management complexity.

Key Results

The simulation results focused on two primary metrics: data integrity and response time.

  • Integrity: Visual comparisons of the "Original database" vs. the "Encrypted database" showed that the proposed algorithm (BS) resulted in no data loss, whereas the MA algorithm showed visible gaps in the data set.
  • Performance (Response Time): The proposed "BS" algorithm outperformed both MA and MB in business processing time (measured in seconds):
  • Data Query: The BS algorithm responded in approximately 8 seconds, compared to ~12s for MA and ~13s for MB.
  • Data Update: The BS algorithm responded in approximately 10 seconds, compared to ~14s for MA and ~13s for MB.

These results indicate that the proposed method is roughly 25-35% faster than the comparison models for standard database operations.

Limitations and Future Work

While the results are promising, the paper leaves a few areas for further exploration:

  • WAN Security: The authors acknowledge that the B/S structure has a "relatively weak security control ability" due to the broad coverage of Wide Area Networks (WANs). While their encryption helps, the broader network security context remains a challenge.
  • Computational Overhead: The paper does not deeply analyze the computational cost of the Lagrange polynomial calculations as the number of user attributes or the depth of the policy tree grows significantly.
  • Real-world Implementation: The study is based on a simulation. Future work would benefit from testing this architecture in a live, high-traffic production environment to see how the B/S server handles massive concurrency.

Summary updated

Abstract

In order to solve the problem that the key of the traditional method is uneven between the user and the authority center, which leads to the long encryption time of a heterogeneous database, a heterogeneous database encryption algorithm based on the B/S structure is proposed. By comparing the advant
BibTeX
@inproceedings{chen2022heterogeneous,
  author = {Jun Chen and Feng Zhang},
  title = {{A Heterogeneous Database Encryption Algorithm Based on B/S Structure}},
  booktitle = {Proceedings of the International Conference on Information Economy, Data Modeling and Cloud Computing, ICIDC 2022, 17-19 June 2022, Qingdao, China},
  year = {2022},
  doi = {10.4108/eai.17-6-2022.2322707},
}
Figure