Build a real-time data warehouse with Flink and Hologres

Seamlessly integrate Flink and Hologres for a unified, real-time data warehouse. This solution streamlines data flow between layers, simplifies architecture, and reduces O&M costs, enabling real-time analytics for scenarios like recommendations and risk control. Key advantages: queryable intermediate data, reusable layers, simple architecture.

Try Now

Intended customers

Enterprises requiring real-time data processing and high-concurrency querying

Users seeking to simplify data architecture and lower O&M costs

Solution advantages

Achieves real-time analytics through Hologres and Flink integration

High performance

Write and update massive datasets in real-time using Flink's built-in Hologres connector. Data is immediately queryable upon writing.

High availability

Hologres achieves resource isolation via primary/secondary deployment and virtual warehouses for read/write separation.

Low O&M

A single pipeline efficiently handles both real-time data serving and OLAP querying.

Architecture and deployment

Build a real-time data warehouse with Flink and Hologres

2400021

The process begins with Flink ingesting source data into Hologres, forming the Operation Data Store (ODS) layer. Flink then leverages the ODS layer's binary logs, transforms the data, and writes these results back to Hologres to create the Data Warehouse Detail (DWD) layer. Following this, Flink consumes DWD binary logs, performs computations, and writes the results to Hologres for the Data Warehouse Service (DWS) layer. Finally, Hologres provides query services for external applications.

90 minutes

$5(This is an estimated cost based on the example resource specifications provided, assuming a 1.5-hour runtime and a data source with three tables, each containing approximately 20 entries. Actual prices may vary due to your selected resource specifications, data volume, runtime, and other factors. The price displayed in the console and on your final bill is definitive.)