Updated on 2024-08-10 GMT+08:00

Introduction to HetuEngine

Introduction to HetuEngine

HetuEngine is a high-performance, interactive SQL analysis and data virtualization engine developed by Huawei. It seamlessly integrates with the big data ecosystem to implement interactive query of massive amounts of data within seconds, and supports cross-source and cross-domain unified data access to enable one-stop SQL convergence analysis in the data lake, between lakes, and between lakehouses.

Basic Concepts to HetuEngine

  • HSBroker: the HetuEngine proxy, which is used for tenant management verification and obtaining the URL for accessing HetuEngine
  • Coordinator: the coordinator of HetuEngine services, responsible for SQL parsing and optimization
  • Worker: responsible for task execution and data processing
  • Connector: an API for HetuEngine to access the database. Through the connector driver, HetuEngine connects to data sources, reads data source metadata, and operates data (adding, deletion, modification, and query).
  • Catalog: the catalog configuration file corresponds to a data source in HetuEngine. A data source can have multiple catalog configurations, which can be configured in the properties file of a data source.
  • Schema: corresponds a schema in a database.
  • Table: corresponds to a table name is a database.

Connection Modes to HetuEngine

Table 1 Connection Modes

Connection Mode

Support for Username/Password Authentication

Support for Keytab Authentication

Support for Cross-Network-Segment Client Access

Prerequisite

HSFabric

Supported

Supported

Supported

  • The node running user services can communicate with the service nodes where HSFabric resides on the HetuEngine server side.
  • Dual-plane network scenarios are supported.
  • Only fixed IP addresses and ports need to be opened for HSFabric.
  • Supported version: MRS 3.1.3 or later.

HSBroker

Supported

Unsupported

Unsupported

  • The node running user services can communicate with the service nodes where HSBroker and Coordinator (randomly distributed in Yarn NodeManger) reside on the HetuEngine server side.
  • A large number of IP addresses and ports need to be opened for coordinators.
  • Supported version: MRS 3.1.0 or later