Use cases
Hologres is a real-time data warehouse compatible with PostgreSQL. It integrates with the Alibaba Cloud big data ecosystem and supports connections to third-party Business Intelligence (BI) tools for real-time and offline data analytics.
Real-time data warehouse
Many analytics pipelines introduce delays: data is batched, transformed overnight, and loaded into a warehouse before anyone can query it. By the time a dashboard refreshes, the underlying data may be hours old.
Hologres eliminates that lag. Stream business data through Realtime Compute for Apache Flink for extract, transform, load (ETL) processing, then query the results in Hologres with sub-second latency. Connect Hologres to BI tools to power live dashboards and reports that always reflect the current state of your business.
Typical workloads:
Real-time dashboards and reports — Display up-to-date business metrics without batch delays.
Real-time monitoring and alerting — Detect anomalies in O&M and application data as they occur, trigger alerts, and support real-time debugging.
Real-time business decisions — Power risk control, recommendation, effect analytics, and model training with current data.
MaxCompute query acceleration
Offline datasets in MaxCompute often hold significant analytical value, but querying them at interactive speeds typically requires long wait times or costly data movement.
Hologres addresses this in two ways: query MaxCompute data directly from Hologres without moving it, or import MaxCompute data into Hologres for faster access. Either way, connect BI tools to Hologres and run interactive analysis on offline datasets.
Typical workloads:
Interactive queries on offline data — Run ad hoc queries against MaxCompute tables at interactive speeds.
Offline reporting — Analyze offline MaxCompute data and generate reports.
Data serving — Expose MaxCompute table data through a RESTful API for online applications.
Federated analytics on real-time and offline data
Not all data needs to live in the same place. Frequently accessed data belongs close to your queries; infrequently accessed data can sit in lower-cost storage.
Partition your data by access frequency: store hot data in Hologres and cold data in MaxCompute. Hologres runs federated queries across both stores. Connect BI tools to run anything from simple lookups to cross-dataset analysis without worrying about where the data lives.
