Data Consistency Simulation Tool for NoSQL Database Systems

TL;DR

This paper introduces a simulation tool for assessing data consistency and violations in NoSQL databases under network partition scenarios, aiding understanding of consistency behaviors.

Contribution

It presents a novel simulation model that measures data consistency and detects violations in NoSQL systems during network partitions.

Findings

Simulation accurately estimates data consistency levels.

Detects and analyzes data consistency violations during network partitions.

Provides insights into the impact of replication and workload on consistency.

Abstract

Various data consistency levels have an important part in the integrity of data and also affect performance especially the data that is replicated many times across or over the cluster. Based on BASE and the theorem of CAP tradeoffs, most systems of NoSQL have more relaxed consistency guarantees than another kind of databases which implement ACID. Most systems of NoSQL gave different methods to adjust a required level of consistency to ensure the minimal numbering of the replicas accepted in each operation. Simulations are always depending on a simplified model and ignore many details and facts about the real system. Therefore, a simulation can only work as an estimation or an explanation vehicle for observed behavior. So to create simulation tool, I have to characterize a model, identify influence factors and simply implement that depending on a (modeled) workload. In this paper, I have a model of simulation to measure the consistency of the data and to detect the data consistency violations in simulated network partition settings. So workloads are needed with the set of users who make requests and then put the results for analysis.