| Studiju veids |
bakalaura akadēmiskās studijas |
| Studiju programmas nosaukums |
Datorsistēmas |
| Nosaukums |
Grafu datu bāzes mērogojamības analīze reāllaika analītikai |
| Nosaukums angļu valodā |
Analysis of Graph Database Scalability for Real-Time Analytics |
| Struktūrvienība |
33000 Datorzinātnes, informācijas tehnoloģijas un enerģētikas fakultāte |
| Darba vadītājs |
Māra Romanovska |
| Recenzents |
Kārlis Berkolds |
| Anotācija |
Real-time analytics over highly interconnected data requires database systems capable of delivering low-latency results under increasing data volumes and concurrent access. Graph databases are well suited to such workloads due to their relationship-centric data model; however, their scalability limits under real-time constraints are not yet fully understood.
This bachelor thesis evaluates the scalability of a graph data model implemented in a graph database under real-time data processing workloads using a single-machine (vertical scaling) architecture. The study focuses on the Neo4j graph database implementing the property graph model and employs a real-world graph dataset to ensure realistic evaluation conditions. A controlled benchmarking methodology is applied to analyse performance across different dataset sizes, memory configurations, query types, and concurrency levels, using latency and throughput as the primary evaluation metrics.
The experimental results show that query structure and computational complexity are the dominant factors affecting real-time performance. Simple traversal-based queries maintain low latency and satisfy interactive response-time requirements, whereas computationally intensive analytical queries exhibit significantly higher latency and limited scalability. Adequate memory allocation improves performance for interactive workloads, while increasing concurrency leads to a rapid increase in tail latency. Based on these findings, practical guidelines are formulated for configuring and using graph databases in real-time analytics scenarios. |
| Atslēgas vārdi |
GRAFU DATUBĀZES, REĀLLAIKA ANALĪTIKA, VERTIKĀLĀ MĒROGOJAMĪBA, NEO4J, VEIKTSPĒJAS SALĪDZINĀŠANA. |
| Atslēgas vārdi angļu valodā |
GRAPH DATABASES, REAL-TIME ANALYTICS, VERTICAL SCALABILITY, NEO4J, PERFORMANCE BENCHMARKING. |
| Valoda |
eng |
| Gads |
2026 |
| Darba augšupielādes datums un laiks |
06.01.2026 02:06:42 |
