Exploring Horizontal Scaling Solutions to Enhance CryptDB Performance under High Concurrency on AWS

This study presents an investigation into the performance of CryptDB, a database system providing practical and provable confidentiality in the face of an adversary who captures all stored data under different cloud-based architectures. The research aims to determine how different configurations impact the scalability and efficiency of CryptDB. The cloud architectures examined include an encrypted environment with no scaling (EENS), an encrypted environment with scaling (EEWS), and an unencrypted data (UE) architecture. A load testing tool, k6, is used to simulate various concurrent user scenarios. The results demonstrate that the proposed EEWS architecture can reduce latency by up to 50%, though it also highlights CryptDB’s limitations, including poor CPU utilization and inadequate documentation. Recommendations are provided for future enterprise projects considering CryptDB. Further research can expand upon this work by considering other alternative tools and their performance under high concurrency.