International Journal of Science, Engineering and TechnologyAuthors: Jaya Ram Menda
Abstract: Modern enterprise systems increasingly demand both sub-millisecond data access and strong durability guarantees to support real-time decision making, regulatory compliance, and continuous availability. No single data store, however, can optimally satisfy these competing requirements across diverse workload profiles, which has led to the adoption of hybrid persistence architectures. This paper examines hybrid persistence strategies that combine Redis for low-latency in-memory access, MongoDB for flexible and horizontally scalable document storage, and PostgreSQL for durable, ACID-compliant transactional consistency. Grounded in the principles of polyglot persistence, the study analyzes widely adopted architectural patterns including cache-aside, write-through, and hybrid read/write routing, illustrating how each pattern balances performance, consistency, and system complexity. The analysis further evaluates trade-offs related to data freshness, failure recovery, operational overhead, and consistency guarantees under partial failures and high concurrency. Drawing on publicly documented case studies and foundational design literature published prior to 2018, the paper proposes a systematic framework for selecting, composing, and governing heterogeneous data stores, enabling scalable enterprise applications to meet stringent latency, reliability, and maintainability requirements simultaneously.
DOI: http://doi.org/10.5281/zenodo.18107916
