Increased Concurrency
QDBase Enhances Bank's Self-Service Analysis Capability
The Challenge
A bank's e-banking self-service analysis system struggled with high concurrency demands, supporting only five concurrent sessions. This limitation stemmed from its reliance on a centralized data warehouse that was near capacity, shared across multiple departments, and was unable to scale further. Attempts to replace the data warehouse or employ conventional relational databases were deemed infeasible due to cost, complexity, and performance constraints.
Efficiency
QDBase is able to increase concurrency from 5 to 50 users per node at a lower cost
Performance
Single query execution reduced from 5 seconds to 2 seconds.
The QDBase Solution
To address these issues, a front-end computation layer using QDBase was introduced.
Improved Query Performance
Single query execution was significantly optimized, reducing the time required from 5 seconds on the data warehouse to just 2 seconds when executed on QDBase.
While the data warehouse relied on a robust 5-node physical cluster with high-end configurations (2x6-core CPUs and 96GB memory per node), QDBase demonstrated superior efficiency using only a virtual machine equipped with a single 2-core CPU and 16GB memory, highlighting the lightweight yet powerful capabilities of the QDBase system.
Increased Concurrency
Before optimization, the system supported only 5 concurrent queries. Post-implementation, each QDBase front-end server comfortably handled 50 concurrent requests, while also maintaining a response time of under 5 seconds.
Enhanced Development Efficiency
The use of QDBase streamlined development processes through its concise scripting and prebuilt functionalities. Tasks that would traditionally require extensive coding and debugging in other environments were completed in less time.
The routing logic, pivotal for segregating hot and cold data processing, was implemented with minimal effort. The flexibility of QDBase enabled continuous refinement of these rules, ensuring optimal performance over time.
Operational Stability
The introduction of QDBase significantly improved the system’s resilience and stability. By reducing the load on the central data warehouse and offloading over 90% of query operations to the front-end computation layer, the overall system operated more smoothly, even during peak usage periods.
QDBase Feature Highlights
- High-Performance Columnar Storage: Enables efficient handling of large-scale data with second-level response times. Effective for processing hot data, reducing computational load on the main data warehouse.
- Concise and Modular Code: Reduces development time and maintenance complexity with its straightforward scripting capabilities. Built-in algorithms for common tasks, eliminating the need for extensive custom coding.
- Flexibility in Integration: Easily integrated with existing workflows by replacing the JDBC driver without altering the core system.
Conclusion
QDBase transformed the bank's self-service analysis system, boosting performance and scalability while minimizing costs and development complexity. By leveraging innovative data routing and high-performance storage, the solution surpassed traditional approaches, setting a new benchmark for efficient financial data analysis.
Read the full in-depth study, with code examples on the SCUDATA Blog
Read about more use cases with real results on our Case Studies page