With the commercialization of open source Hapdoop, Map/reduce, Spark, HDFS, HBase and other technologies, large data management technology has been developed by leaps and bounds. How to objectively compare different data management systems, that is, the choice of large data test benchmark, has become an important research topic.
The Transactional Energy Management Committee (transactionprocessing performance COUNCIL,TPC) is currently the most well-known standard organization for evaluating benchmarks for non-profit data management systems. It defines multiple sets of standard test sets for objectively and reproducible evaluation of database performance. Over the past more than 20 years, the agency has published a variety of database benchmarking benchmarks, such as Tpc-a, tpc-d, Tpc-h and Tpc-ds. The Decisionsupport (DS) subset, Tpc-ds, is a standard set of SQL tests for evaluating decision support systems (or data warehouses). This test set contains a large data set of statistics, report generation, online query, data mining and other complex applications, testing data and values are skewed, consistent with the real data. It can be said that Tpc-ds is very close to the real scene of a test set, but also a more difficult test set.
TPC is comprised of more than 10 server vendors (Huawei is the only Chinese company in the organization) involved in the development of standard specification, performance and price metrics for business application benchmarks (Benchmark), and manages the release of test results. TPC benchmark test results are the server class equipment performance of the core technical indicators.