Quick BI, Alibaba Cloud's cloud business intelligence (BI) software, is built on the Apsara distributed operating system. It supports both Software as a Service (SaaS) and private deployment, catering to consumption-based BI across various scenarios, terminals, and industries. This topic provides an insight into the acceleration engine powering Quick BI.
Developed on Alibaba Cloud's scalable architecture, Quick BI offers traditional BI features, including visual analytics, reporting, and self-service analytics. It also boasts enterprise-grade security and integrates seamlessly with mobile devices and third-party systems.
At its core, Quick BI utilizes a proprietary computing kernel as an acceleration engine, capable of aggregating and analyzing billions of records from SaaS services hosted on Alibaba Cloud in just 0.5 seconds. Alibaba Cloud's computing resources are scalable, allowing the addition of servers to enhance the engine's data analytics and computing power.
Why design a new Quick engine
With the progression of digital transformation, enterprises increasingly recognize the importance of data-driven decision-making and analysis. BI services are widely used by a diverse range of industries and business sizes to create reports, dashboards, and BI portals for informed decision-making and analysis.
However, when utilizing BI services for data analysis, the following scenarios can lead to issues if data processing is slow:
Presenting a report to your boss can be frustrating if the data loads slowly or fails to load, resulting in a poor user experience.
When an analyst or business department member drags a metric for exploratory analysis, delayed results can significantly hinder work efficiency and disrupt the analysis process.
Slow data processing detracts from the user experience. To address these issues, BI services require a robust big data processing architecture that supports scaling out to accommodate growing data volumes and computational demands.
The Quick engine is strategically positioned between data sources and datasets, processing queries from upper-layer applications and directing them to the appropriate data sources.
Multi-mode BI computing engine
Quick BI's acceleration engine serves as the computing foundation, offering a multi-mode BI computing engine that supports various computing modes. These include direct database connection, real-time acceleration, query result caching, and dimension value acceleration, ensuring efficient computing solutions tailored to different scenarios.
Quick BI's operation involves data sources, datasets, and data works. A data source establishes the underlying database connection, while a dataset models tables within a data source, allowing for table associations and data type modeling. This transforms tables into data objects ready for upper-layer applications like dashboards, workbooks, and ad hoc queries.
The Quick engine offers the following computing modes for various scenarios:
Direct connection mode: This mode executes computing tasks directly on the connected database or data warehouse, supporting a wide array of data sources and accessible to all users. It is ideal for scenarios where the underlying computing resources can handle the query load.
Configure the Quick engine: This mode extracts data from user databases or data warehouses into the high-performance columnar storage engine of the Quick engine. It supports both full and incremental modes. Analysis computing loads run directly in the Quick engine, fully utilizing its performance while reducing the burden on user data warehouses. This mode is available to Pro and Professional Edition users and is particularly suitable for enterprises without independent data warehouses or those with overloaded data warehouses.
Query caching: Accessible to all users, this mode caches temporary query results when reports or dashboards are accessed. If identical queries are made within the cache validity period, cached results are retrieved, speeding up response times and reducing the need for repeated calculations. It is particularly effective for scenarios with frequent, repeated queries, such as visual displays.
Dimension value acceleration: Also available to all users, this mode leverages direct connection and dimension table configurations to expedite high-frequency, time-consuming dimension field queries. It performs these queries on the database's dimension table rather than the original detail table, offering quick results and conserving computing resources in scenarios like ad hoc analysis and dimension value queries in query controls.