Baidu officially opened yesterdayPaloThis is a Baidu based, interactive SQL data warehouse based on MPP, which is mainly used for solving reports and multidimensional analysis.
Palo mainly integrates Google, Mesa and Cloudera Impala technology. Unlike other popular SQL-on-Hadoop systems, Palo is designed to be a single, tightly coupled system that does not rely on other systems.
Palo not only provides high concurrency and low latency query performance, but also provides high throughput ad-hoc analysis queries. It also provides bulk data loading, as well as near real-time small volume data loading.
Palo has high availability, reliability, fault tolerance and scalability. Its main features are simple (development, deployment and use) and meet many data service requirements in a single system.
The implementation of Palo consists of two daemons: the FE and the BE. The following illustration gives an overview of the architecture and usage:
Palo's name is OLAP. Write it upside down