Researchers: Mohammed Jafar Suleiman Al-Nahdi, Abdullah Ahmed Mahfoudh Sowailem
Supervisor: Dr. Abdullah Al-Dhibani
Enhancing Digital Record Integrity Using Immutable Storage Techniques in Relational DBMS
Data Integrity is considered one of the fundamental pillars of information security, especially in financial, governmental, and medical systems. However, traditional Relational Database Management Systems (RDBMS) face a critical security gap related to Insider Threats.
In current infrastructures, Database Administrators (DBAs) or accounts with high privileges (Superusers) have absolute ability to execute UPDATE or DELETE operations on any record within tables, regardless of the Access Control Policies applied to ordinary users. This privilege model creates a Single Point of Failure, where any authorized party, whether motivated by malice or due to account compromise, can manipulate historical records without leaving a clear trace. This undermines trust in the integrity of stored data and threatens compliance with regulatory standards.
This study aims to design and develop a software framework that ensures Immutability for data in specific tables within a database, regardless of the privilege level granted to the user attempting access.
The proposed solution is based on the Append-Only Log principle, where modifying or deleting records is strictly prohibited once written to specific tables. The solution does not rely on bypassable software privilege policies, but rather on fundamental modifications to the Database Engine itself to enforce constraints at the Storage Engine Level.
Given the flexibility of the PostgreSQL DBMS as an open-source system, a Database Extension development methodology was adopted to modify the Kernel behavior. The methodology includes the following steps:
A software module will be developed to add new keywords to the Structured Query Language (SQL). Instead of the traditional CREATE TABLE command, the system will introduce a new command: CREATE IMMUTABLE_TABLE.
- Academic Justification: This allows defining the Data Contract at the design phase, clarifying the developer's intention that this table is dedicated to permanent, immutable records.
When the new command is invoked, the system internally performs the following:
- Creates the table with strict Constraints preventing UPDATE and DELETE operations.
- Modifies Storage Parameters to ensure that Pages allocated for this table are treated as Read-Only after initial writing.
- Intercepts any attempt to directly access data files (WAL files) to modify records.
If data correction is required, the system does not allow direct modification. Instead, it mandates inserting a new record referencing the previous record (Linked List Structure), thereby preserving the complete Audit Trail.
To enhance the scientific and practical value of the thesis, the following ideas are proposed for integration within the study scope:
- Idea: Preventing modification programmatically is insufficient; it must be proven mathematically. It is proposed that the system automatically generates a Hash for each new record and links it to the hash of the previous record (similar to Blockchain but within the database).
- Benefit: This allows immediate detection of any potential tampering in physical storage files, even if the database application layer is bypassed.
- Idea: Conduct a precise comparative study between traditional tables and immutable tables regarding Write Throughput and consumed storage size.
- Benefit: Determine whether the Integrity Cost is acceptable in high-performance systems (e.g., High-Frequency Trading Systems).
- Idea: Discuss the issue of the Right to be Forgotten in the context of immutable tables.
- Proposed Solution: Develop a Cryptographic Shredding mechanism, where data is effectively deleted by removing its specific encryption keys without needing to modify the record itself, achieving a balance between integrity and privacy.
- Idea: Allow trusted external parties to verify record integrity (Verification) without granting them access to the data content (Zero-Knowledge Proof).
- Benefit: Enhances transparency in governmental systems or supply chains.
This study is expected to prove that integrating Immutable Storage Techniques within open-source Relational Database Systems provides an effective additional security layer against Insider Threats, while maintaining system efficiency. Furthermore, the thesis will present a Prototype for a PostgreSQL extension that can be adopted as a security standard in sensitive systems.