Enterprise digital analytics becomes diversified. It involves real-time business intelligence (BI) decisions, real-time reports, data extract-transform-load (ETL), data cleansing, and AI analysis.

Traditional data warehousing solutions combine multiple sets of databases and big data tools, and use their advantages in different analysis scenarios. This leads to data redundancy and increasing management costs of multiple heterogeneous systems.

To improve database warehouses, the following issues must be resolved:
  • How to better support interactive analytics in database scenarios and complex batch computing in big data scenarios.
  • How to provide an end-to-end service solution for hybrid loads.

AnalyticDB for MySQL is a new generation of cloud native data warehouse that provides an end-to-end data warehouse service.

Hybrid computing engine

AnalyticDB for MySQL supports interactive and batch computing modes. You can analyze data in real time while low latency is ensured in the interactive query scenario, and achieve high throughput for batch computing of big data in the complex offline computing scenario.

  • In interactive mode, AnalyticDB for MySQL utilizes a massively parallel processing (MPP) architecture that schedules all query tasks and uses pipeline streaming computing. This mode is ideal for the interactive analysis scenario that requires low latency.
  • In batch mode, AnalyticDB for MySQL utilizes a batching computing architecture that uses directed acyclic graphs (DAGs) to divide and schedule tasks in batches. This can enable large amounts of data to be computed by using limited resources. Data can be written to disks after computing. This mode is ideal for the complex analysis scenario that involves large amounts of data and requires high throughput.

Hybrid load isolation by resource pool

AnalyticDB for MySQL in elastic mode allows you to isolate multiple tenants by resource pool. You can use a single cluster to meet resource requirements in different analysis scenarios.

Resources in resource pools are physically isolated. You can manage resources in these pools, query the engine mode, and manage related configuration information.

Query traffic is automatically routed to corresponding resource pools based on the account that you use. This can minimize interruptions to your business. You can also use hints to route query traffic to corresponding resource pools.