ISSN 2996-8223
International Journal of Computer Science and Technology | Vol. 3, No. 9, September 2022 | pp. 1–8
DOI: 10.46882/2022/IJCST/000212
Article Type: Original Research Paper
Title: Minimizing Query Delays in Geographically Replicated NoSQL Systems via Dynamic Consensus Scaling
Names of Authors: Elena Rostova¹, Jean-Luc Picard²
Authors’ Affiliations: ¹Data Engineering Laboratory, NordTech University, Oslo, Norway; ²Distributed Systems Architecture Division, SynthData Systems, Paris, France
Abstract: Maintaining strict transactional consistency across globally distributed database systems typically leads to high query latency and performance bottlenecks. This research proposes an optimized database replication framework that uses an adaptive variant of the Raft consensus algorithm. The algorithm dynamically measures regional round-trip network performance and modifies quorum execution targets based on localized traffic requirements. When operational volatility is minimal, the system relaxes global synchronization states down to localized clusters. It returns to full multi-region confirmation phases exclusively during high write-concurrency periods. We deployed the framework across a distributed cloud testbed containing 12 distinct geographic nodes to evaluate performance. The experimental findings indicate a 42.1% reduction in mean write latency compared to standard, non-adaptive Raft configurations. The transactional throughout achieved a sustainable capacity of 18,500 operations per second under complex read-to-write workloads. This dynamic adjustment methodology minimizes global operational staleness without compromising overall transactional durability metrics.
Keywords: NoSQL Databases, Raft Consensus, Data Replication, Query Latency, Distributed Systems, Transactional Consistency
Manuscript Timeline: Received: May 15, 2022; Revised: June 24, 2022; Accepted: July 19, 2022; Published: September 05, 2022
Subscribe to read the full article: https://internationalscholarsjournals.org/subscribe-to-read