E-MapReduce:Kyuubi

Features

• Multi-tenancy: Kyuubi provides end-to-end multi-tenancy for resource acquisition and access to data or metadata by using a unified authentication or authorization layer.

• High availability: Kyuubi supports load balancing by using ZooKeeper. This ensures an enterprise-level high availability and an unlimited high client concurrency.

• Multiple workloads: Kyuubi can easily support multiple workloads by using one platform, one copy of data, and one SQL interface.

Scenarios

• Interactive analytics: Kyuubi helps build an enterprise-level analytic platform for visualized interactive analytics on big data. The platform supports common computing frameworks. Kyuubi supports Java Database Connectivity (JDBC) and Open Database Connectivity (ODBC) interfaces. You can use SQL or business intelligence (BI) tools to access Kyuubi and efficiently perform queries. Kyuubi caches background engine instances at the user level to ensure computing resource sharing and quick response. This way, large amounts of data can be queried in parallel and the query results can be quickly returned.

• Batch processing: Kyuubi provides an SQL interface for batch processing, especially for large-scale extract, transform, and load (ETL) processes. Kyuubi and its engines support independent storage and a number of data sources. Kyuubi isolates background engine instances at the connection level to improve computing resource isolation and stability.

Comparison among Kyuubi, Livy, and Spark Thrift Server

References

• The Share Level Of Kyuubi Engines

• The TTL Of Kyuubi Engines