MANAGING CONFLICT RATE REDUCTION IN SCALABLE DISTRIBUTED DATABASE SYSTEMS

Authors

Vipul Kumar Bondugula

Abstract

Databases are foundational components in modern computing systems, used to store, manage, and retrieve structured data efficiently. As data volumes and access requirements grow, traditional centralized databases often struggle to meet performance, availability, and fault tolerance demands. To address these limitations, distributed databases have emerged as a scalable solution, spreading data across multiple nodes or geographic locations. This architecture improves system resilience and enables faster access to data by colocating it closer to users. However, distributed databases also introduce new complexities, particularly in maintaining consistency across multiple nodes and managing concurrent access from numerous transactions. Concurrency control is a critical aspect of distributed database systems, ensuring that multiple transactions can execute simultaneously without compromising data integrity. In high-traffic environments, concurrent transactions often access and modify the same data items, leading to potential conflicts. These conflicts must be detected and resolved efficiently to preserve the correctness of operations. The conflict rate, which indicates how frequently transactions interfere with each other, is a key performance metric in such systems. High conflict rates result in increased transaction aborts and retries, leading to reduced throughput and higher latency. Snapshot Isolation (SI) is a widely used concurrency control mechanism in distributed databases. SI allows transactions to operate on a consistent snapshot of the database taken at their start time, avoiding read-write conflicts. While SI is effective at eliminating many anomalies and providing a user-friendly isolation level, it struggles under certain conditions. One of the primary challenges faced by SI is write skew or write-write conflicts, which occur when multiple transactions modify overlapping sets of data concurrently. SI does not prevent such conflicts effectively, leading to anomalies that can compromise consistency. As a result, systems relying heavily on SI may observe a growing number of conflicts and transaction aborts, particularly in write-intensive workloads. This paper addresses on conflicts rate reduction management by using multi version concurrency control.

Keywords

MANAGING CONFLICT RATE REDUCTION IN SCALABLE DISTRIBUTED DATABASE SYSTEMS. Vipul Kumar Bondugula. 2023. IJIRCT, Volume 9, Issue 1. Pages 1-23. https://www.ijirct.org/viewPaper.php?paperId=2505017