RDBMS: Dataset & Load Generator

2nd September 2021 at 6:08pm

来源自 Materialize 的 文章

For this demo, Materialize uses a custom benchmark called chBench that is really just a concatenation of two well-known database benchmarking tools, TPC-C and TPC-H.

TPC-C is an industry-standard benchmark for transactional workloads, meant to test a database’s scalability.

TPC-H is from the same group that developed TPC-C, and is meant to test the capabilities of online analytic platforms by performing complex analytic style queries. This includes large aggregations, many groupings, and complex multi-way joins.

CH-benCHmark brings together a TPC-C-like dataset with TPC-H’s analytical queries. This is a great approximation for how many businesses perform OLAP queries over OLTP data.

大意是说,TPC-C 能产生一个数据集,并产生重事务的大量请求。TPC-H 能产生大量复杂查询的请求。CH-benCHmark 将两者整合。

应用场景

  • 验证一些数据同步过程是否保持了一致性,比如把 RDBMS 的改动同步到数据仓库中(如 Hive 表)