Evaluating Database Encryption: a Comparative Study of PostgreSQL Implementations With EnterpriseDB TDE and Cybertec TDE

Abstract

In the digital era, protecting sensitive data has become an urgent necessity as incidents of identity theft and information breaches continue to rise. Transparent Data Encryption (TDE) offers a solution by enabling storage-level encryption without requiring changes to applications; however, PostgreSQL does not natively support TDE. Vendors such as EnterpriseDB and Cybertec have developed TDE implementations that can be integrated to address this security need. This study aims to explore the implementation of TDE on PostgreSQL and evaluate its impact on system performance and data security. An experimental approach was employed using the pgbench benchmarking tool. Performance tests were conducted on databases of varying sizes—from 50 to 5,000,000 rows under both low‑load and high‑load conditions. The primary metrics were transactions per second (TPS) and average transaction latency. Test results indicate that implementing TDE significantly enhances resistance to unauthorized access but incurs a notable performance penalty. The Cybertec non-TDE solution achieved the highest TPS and lowest latency, whereas Cybertec with TDE experienced an encryption overhead that reduced throughput by approximately 25–35%. Meanwhile, the EnterpriseDB non-TDE configuration demonstrated more stable performance, exhibiting a smaller standard deviation compared to the TDE implementation. In conclusion, adopting TDE on PostgreSQL substantially increases data protection, yet organizations must weigh this security benefit against the resultant performance decline. Future recommendations include optimizing encryption configurations and investigating hardware‑acceleration techniques to minimize overhead.