International Journal on Science and Technology

E-ISSN: 2229-7677     Impact Factor: 9.88

A Widely Indexed Open Access Peer Reviewed Multidisciplinary Bi-monthly Scholarly International Journal

Call for Paper Volume 16 Issue 2 April-June 2025 Submit your research before last 3 days of June to publish your research paper in the issue of April-June.

Enhancing Scalability with Optimized Replication in Distributed Architectures

Author(s) Kanagalakshmi Murugan
Country United States
Abstract Viewstamped Replication (VR) is a consensus protocol widely used in distributed systems to ensure consistency and fault tolerance among replicated services. It operates through a quorum-based mechanism, where a majority of nodes must agree before a client’s request can be committed. While VR is reliable and effective in maintaining data consistency across nodes, one significant drawback observed in its operation is the high replication time, particularly as the number of participating nodes increases. This issue becomes more evident in larger clusters, where the overhead of communication and coordination among nodes grows substantially. Replication time refers to the duration taken by the protocol to replicate and confirm a client’s request across all required nodes. In VR, each operation involves multiple rounds of messaging between the primary (leader) and backup nodes. As the system scales, the number of these interactions increases, adding delay to the overall replication process. For instance, in a system with 3 nodes, the replication time is relatively low. However, as the cluster expands to 5, 7, 9, or 11 nodes, the time required for replication rises significantly, following a clear upward trend. This increase is due to the need for reaching a quorum and synchronizing responses from more nodes, which introduces network latency, processing delay, and potential queuing overheads. The high replication time of VR can lead to reduced system throughput and slower response times for client operations, making it less suitable for performance-critical applications or large-scale distributed environments. Furthermore, in real-time systems where low latency is crucial, such delays can hinder the system’s ability to meet strict timing requirements. This limitation presents a scalability challenge for VR, as the benefits of adding more nodes for fault tolerance and availability come at the cost of increased latency. Viewstamped Replication (VR) is a reliable consensus protocol for distributed systems, ensuring consistency and fault tolerance. However, its replication time increases significantly as the number of nodes grows. This happens because VR relies on a quorum-based approach, requiring multiple interactions and acknowledgments from a majority of nodes for each operation. Without addressing the replication delay, systems relying on VR may struggle to maintain high performance as they grow, making this an important area of concern in the field of distributed computing. This paper addresses the replication time issue in VR by using the ZAB replication time.
Field Engineering
Published In Volume 14, Issue 3, July-September 2023
Published On 2023-08-04
DOI https://doi.org/10.71097/IJSAT.v14.i3.6272
Short DOI https://doi.org/g9p7mq

Share this