Star Schema Benchmark

This document mainly introduces how to pass the preliminary performance test of the SSB process in Doris.

Please refer to the official document (opens new window) to install and deploy Doris to obtain a normal running Doris cluster ( Contain at least 1 FE, 1 BE).

The scripts involved in the following documents are all stored under in the Doris code base.

Execute the following script to download and compile the tool.

After the installation is successful, the dbgen binary file will be generated in the ssb-dbgen/ directory.

2. Generate SSB test set

Execute the following script to generate the SSB data set:

Under the -s 100 parameter, the generated data set size is:

1. Import data

   1. Prepare the ‘doris-cluster.conf’ file.

      Before calling the load script, you need to write the FE’s ip port and other information in the doris-cluster.conf file.

      ‘doris-cluster.conf’ in the same directory as load-dimension-data.sh.

      The contents of the file include FE’s ip, HTTP port, user name, password and the DB name of the data to be loaded:

   2. Load 4 dimension table data (customer, part, supplier and date)

      Because the data volume of these 4 dimension tables is small, and the load is simpler, we use the following command to load the data of these 4 tables first:

      sh load-dimension-data.sh

   3. Load the fact table lineorder.

      Load the lineorder table data with the following command:

      sh load-fact-data.sh -c 5

2. Check the loaded data

   The amount of data should be the same as the number of rows of generated data.

There are 4 groups of 14 SQL in the SSB test set. The query statement is in the directory.

The following test report is based on Doris branch-0.15 (opens new window) branch code test, for reference only. (Update time: October 25, 2021)

1. Hardware environment

   • CPU: 96core, Intel(R) Xeon(R) Gold 6271C CPU @ 2.60GHz
   • Memory: 384GB
   • Hard disk: 1 HDD
   • Network card: 10 Gigabit network card

2. Data set

3. Test results