Benchmark.md

Benchmark Tool

1. Basic Overview

IoT-benchmark is a time-series database benchmarking tool developed in Java for big data environments. It was developed and open-sourced by the School of Software, Tsinghua University. The tool is user-friendly, supports various write and query methods, allows storing test information and results for further queries or analysis, and integrates with Tableau for visualizing test results.

Figure 1-1 illustrates the test benchmark process and its extended functionalities, all of which can be streamlined by IoT-benchmark. It supports a variety of workloads, including write-only, read-only, and mixed write-and-read operations. Additionally, it offers software and hardware system monitoring, performance metric measurement, automated database initialization, test data analysis, and system parameter optimization.

Figure 1-1 IoT-benchmark Test Benchmark Process

IoT-benchmark adopts the modular design concept of the YCSB test tool, which separates workload generation, performance measurement, and database interface components. Its modular structure is illustrated in Figure 1-2. Unlike YCSB-based testing tools, IoT-benchmark introduces a system monitoring module that supports the persistence of both test data and system metrics. It also includes load-testing functionalities specifically designed for time-series data scenarios, such as batch writes and multiple out-of-order data insertion modes for IoT environments.

Figure 1-2 IoT-benchmark Modular Design

Supported Databases

Currently, IoT-benchmark supports the following time series databases, versions and connection methods:

Database	Version	Connection mmethod
IoTDB	v1.x v2.x	JDBC, SessionByTablet, SessionByRecord, SessionByRecords
InfluxDB	v1.x v2.x	SDK
TimescaleDB	--	JDBC
OpenTSDB	--	HTTP Request
QuestDB	v6.0.7	JDBC
TDengine	v2.2.0.2	JDBC
VictoriaMetrics	v1.64.0	HTTP Request
KairosDB	--	HTTP Request

2. Installation and Operation

2.1 Prerequisites

1. Java 8
2. Maven 3.6+
3. The corresponding appropriate version of the database, such as Apache IoTDB 2.0

2.2 How to Obtain

• Binary package: Visit https://github.com/thulab/iot-benchmark/releases to download the installation package. Extract the compressed file into a desired folder for use.

• Source Code Compilation (for Apache IoTDB 2.0 testing):

  • Compile the latest IoTDB Session package: Download the IoTDB source code from https://github.com/apache/iotdb/tree/rc/2.0.5 and run the following command in the root directory to compile the latest IoTDB Session package:

     mvn clean package install -pl session -am -DskipTests

  • Compile the IoT-benchmark test package: Download the source code from https://github.com/thulab/iot-benchmark and run the following command in the root directory to compile the Apache IoTDB 2.0 test package:.

     mvn clean package install -pl iotdb-2.0 -am -DskipTests

  • The compiled test package will be located at:

  ./iotdb-2.0/target/iotdb-2.0-0.0.1/iotdb-2.0-0.0.1

2.3 Test Package Structure

The directory structure of the test package is shown below. The test configuration file is conf/config.properties, and the test startup scripts are benchmark.sh (Linux & MacOS) and benchmark.bat (Windows). The detailed usage of the files is shown in the table below.

-rw-r--r--. 1 root root  2881 Jan  10 01:36 benchmark.bat
-rwxr-xr-x. 1 root root   314 Jan  10 01:36 benchmark.sh
drwxr-xr-x. 2 root root    24 Jan  10 01:36 bin
-rwxr-xr-x. 1 root root  1140 Jan  10 01:36 cli-benchmark.sh
drwxr-xr-x. 2 root root   107 Jan  10 01:36 conf
drwxr-xr-x. 2 root root  4096 Jan  10 01:38 lib
-rw-r--r--. 1 root root 11357 Jan  10 01:36 LICENSE
-rwxr-xr-x. 1 root root   939 Jan  10 01:36 rep-benchmark.sh
-rw-r--r--. 1 root root    14 Jan  10 01:36 routine

Name	File	Usage
benchmark.bat	-	Startup script on Windows
benchmark.sh	-	Startup script on Linux/Mac
bin	startup.sh	Initialization script folder
conf	config.properties	Test scenario configuration file
lib	-	Dependency library
LICENSE	-	License file
cli-benchmark.sh	-	One-click startup script
routine	-	Automatic execution of multiple test configurations
rep-benchmark.sh	-	Automatic execution of multiple test scripts

2.4 Execution of Tests

1. Modify the configuration file (conf/config.properties) according to test requirements. For example, to test Apache IoTDB 2.0, set the following parameter:

      DB_SWITCH=IoTDB-200-SESSION_BY_TABLET

2. Ensure the target time-series database is running.

3. Start IoT-benchmark to execute the test. Monitor the status of both the target database and IoT-benchmark during execution.

4. Upon completion, review the results and analyze the test process.

2.5 Results Interpretation

All test log files are stored in the logs folder, while test results are saved in the data/csvOutput folder. For example, the following result matrix illustrates the test outcome:

• Result Matrix:
  • OkOperation: Number of successful operations.
  • OkPoint: Number of successfully written points (for write operations) or successfully queried points (for query operations).
  • FailOperation: Number of failed operations.
  • FailPoint: Number of failed write points.
• Latency (ms) Matrix:
  • AVG: Average operation latency.
  • MIN: Minimum operation latency.
  • Pn: Quantile values of the overall operation distribution (e.g., P25 represents the 25th percentile, or lower quartile).

3. Main Parameters

3.1 IoTDB Service Model

The IoTDB_DIALECT_MODE parameter supports two modes: tree and table. The default value is tree.

• For IoTDB 2.0 and later versions, the IoTDB_DIALECT_MODE parameter must be specified, and only one mode can be set for each IoTDB instance.
• IoTDB_DIALECT_MODE = tree:
  • The number of devices must be greater than or equal to the number of databases.

3.2 Working Mode

The BENCHMARK_WORK_MODE parameter supports four operational modes:

1. General Test Mode (testWithDefaultPath): Configured via the OPERATION_PROPORTION parameter to support write-only, read-only, and mixed read-write operations.
2. Data Generation Mode (generateDataMode): Generates a reusable dataset, which is saved to FILE_PATH for subsequent use in the correctness write and correctness query modes.
3. Single Database Correctness Write Mode (verificationWriteMode): Verifies the correctness of dataset writing by writing the dataset generated in data generation mode. This mode supports only IoTDB v1.0+ and InfluxDB v1.x.
4. Single Database Correctness Query Mode (verificationQueryMode): Verifies the correctness of dataset queries after using the correctness write mode. This mode supports only IoTDB v1.0+ and InfluxDB v1.x.

Mode configurations are shown in the following below:

Mode name	BENCHMARK_WORK_MODE	Description	Required Configuration
General test mode	testWithDefaultPath	Supports multiple read and write mixed load operations.	OPERATION_PROPORTION
Generate data mode	generateDataMode	Generates datasets recognizable by IoT-benchmark.	FILE_PATH and DATA_SET
Single database correctness write mode	verificationWriteMode	Writes datasets for correctness verification.	FILE_PATH and DATA_SET
Single database correctness query mode	verificationQueryMode	Queries datasets to verify correctness.	FILE_PATH and DATA_SET

3.3 Server Connection Information

Once the working mode is specified, the following parameters must be configured to inform IoT-benchmark of the target time-series database:

Parameter	Type	Example	Description
DB_SWITCH	String	IoTDB-200-SESSION_BY_TABLET	Specifies the type of time-series database under test.
HOST	String	127.0.0.1	Network address of the target time-series database.
PORT	Integer	6667	Network port of the target time-series database.
USERNAME	String	root	Login username for the time-series database.
PASSWORD	String	root	Password for the database login user.
DB_NAME	String	test	Name of the target time-series database.
TOKEN	String	-	Authentication token (used for InfluxDB 2.0).

3.4 Write Scenario Parameters

Parameter	Type	Example	Description
CLIENT_NUMBER	Integer	100	Total number of clients used for writing.
GROUP_NUMBER	Integer	20	Number of databases (only applicable for IoTDB).
DEVICE_NUMBER	Integer	100	Total number of devices.
SENSOR_NUMBER	Integer	300	Total number of sensors per device. (Control the number of attribute columns if you use the IoTDB table model)
INSERT_DATATYPE_PROPORTION	String	1:1:1:1:1:1:0:0:0:0	Ratio of data types: BOOLEAN:INT32:INT64:FLOAT:DOUBLE:TEXT:STRING:BLOB:TIMESTAMP:DATE.
POINT_STEP	Integer	1000	Time interval (in ms) between generated data points.
OP_MIN_INTERVAL	Integer	0	Minimum execution interval for operations (ms): if the operation takes more than the value, the next one will be executed immediately, otherwise wait (OP_MIN_INTERVAL - actual execution time) ms; if it is 0, the parameter is not effective; if it is -1, its value is consistent with POINT_STEP
IS_OUT_OF_ORDER	Boolean	false	Specifies whether to write data out of order.
OUT_OF_ORDER_RATIO	Floating point number	0.3	Proportion of out-of-order data.
BATCH_SIZE_PER_WRITE	Integer	1	Number of data rows written per batch.
START_TIME	Time	2022-10-30T00:00:00+08:00	Start timestamp for data generation.
LOOP	Integer	86400	Total number of write operations: Each type of operation will be divided according to the proportion defined by OPERATION_PROPORTION
OPERATION_PROPORTION	Character	1:0:0:0:0:0:0:0:0:0:0	Ratio of operation types (write:Q1:Q2:...:Q10).

3.5 Query Scenario Parameters

Parameter	Type	Example	Description
QUERY_DEVICE_NUM	Integer	2	Number of devices involved in each query statement.
QUERY_SENSOR_NUM	Integer	2	Number of sensors involved in each query statement.
QUERY_AGGREGATE_FUN	Character	count	Aggregate functions used in queries (COUNT, AVG, SUM, etc.).
STEP_SIZE	Integer	1	Time interval step for time filter conditions.
QUERY_INTERVAL	Integer	250000	Time interval between query start and end times.
QUERY_LOWER_VALUE	Integer	-5	Threshold for conditional queries (WHERE value > QUERY_LOWER_VALUE).
GROUP_BY_TIME_UNIT	Integer	20000	The size of the group in the GROUP BY statement
LOOP	Integer	10	Total number of query operations: Each type of operation will be divided according to the proportion defined by OPERATION_PROPORTION
OPERATION_PROPORTION	Character	0:0:0:0:0:0:0:0:0:0:1	Ratio of operation types (write:Q1:Q2:...:Q10).

3.6 Query Types and Example SQL

Number	Query Type	IoTDB Sample SQL
Q1	Precise Point Query	select v1 from root.db.d1 where time = ?
Q2	Time Range Query	select v1 from root.db.d1 where time > ? and time < ?
Q3	Time Range with Value Filter	select v1 from root.db.d1 where time > ? and time < ? and v1 > ?
Q4	Time Range Aggregation Query	select count(v1) from root.db.d1 where and time > ? and time < ?
Q5	Full-Time Range with Filtering	select count(v1) from root.db.d1 where v1 > ?
Q6	Range Aggregation with Filter	select count(v1) from root.db.d1 where v1 > ? and time > ? and time < ?
Q7	Time Grouping Aggregation	select count(v1) from root.db.d1 group by ([?, ?), ?, ?)
Q8	Latest Point Query	select last v1 from root.db.d1
Q9	Descending Range Query	select v1 from root.sg.d1 where time > ? and time < ? order by time desc
Q10	Descending Range with Filter	select v1 from root.sg.d1 where time > ? and time < ? and v1 > ? order by time desc

3.7 Test process and test result persistence

IoT-benchmark currently supports persisting the test process and test results through configuration parameters.

Parameter	Type	Example	Description
TEST_DATA_PERSISTENCE	String	None	Specifies the result persistence method. Options: None, IoTDB, MySQL, CSV.
RECORD_SPLIT	Boolean	true	Whether to split results into multiple records. (Not supported by IoTDB currently.)
RECORD_SPLIT_MAX_LINE	Integer	10000000	Maximum number of rows per record (10 million rows per database table or CSV file).
TEST_DATA_STORE_IP	String	127.0.0.1	IP address of the database for result storage.
TEST_DATA_STORE_PORT	Integer	6667	Port number of the output database.
TEST_DATA_STORE_DB	String	result	Name of the output database.
TEST_DATA_STORE_USER	String	root	Username for accessing the output database.
TEST_DATA_STORE_PW	String	root	Password for accessing the output database.

Result Persistence Details

• CSV Mode: If TEST_DATA_PERSISTENCE is set to CSV, a data folder is generated in the IoT-benchmark root directory during and after test execution. This folder contains:
  • csv folder: Records the test process.
  • csvOutput folder: Stores the test results.
• MySQL Mode: If TEST_DATA_PERSISTENCE is set to MySQL, IoT-benchmark creates the following tables in the specified MySQL database:
  • Test Process Table:
    1. Created before the test starts.
    2. Named as: testWithDefaultPath_<database_name>_<remarks>_<test_start_time>.
  • Configuration Table:
    1. Named CONFIG.
    2. Stores the test configuration.
    3. Created if it does not exist.
  • Final Result Table:
    1. Named FINAL_RESULT.
    2. Stores the test results after test completion.
    3. Created if it does not exist.

3.8 Automation Script

One-Click Script Startup

The cli-benchmark.sh script allows one-click startup of IoTDB, IoTDB Benchmark monitoring, and IoTDB Benchmark testing. However, please note that this script will clear all existing data in IoTDB during startup, so use it with caution.

Steps to Run:

1. Edit the IOTDB_HOME parameter in cli-benchmark.sh to the local IoTDB directory.
2. Start the test by running the following command:

> ./cli-benchmark.sh

1. After the test completes:
2. Check test-related logs in the logs folder.
3. Check monitoring-related logs in the server-logs folder.

Automatic Execution of Multiple Tests

Single tests are often insufficient without comparative results. Therefore, IoT-benchmark provides an interface for executing multiple tests in sequence.

1. Routine Configuration: Each line in the routine file specifies the parameters that change for each test. For example:

      LOOP=10 DEVICE_NUMBER=100 TEST
      LOOP=20 DEVICE_NUMBER=50 TEST
      LOOP=50 DEVICE_NUMBER=20 TEST
   

In this example, three tests will run sequentially with LOOP values of 10, 20, and 50.

Then the test process with 3 LOOP parameters of 10, 20, and 50 is executed in sequence.

Important Notes:

• Multiple parameters can be changed in each test using the format:

    LOOP=20 DEVICE_NUMBER=10 TEST

• Avoid unnecessary spaces.

• The TEST keyword marks the start of a new test.

• Changed parameters persist across subsequent tests unless explicitly reset.

2. Start the Test: After configuring the routine file, start multi-test execution using the following command

      > ./rep-benchmark.sh

Test results will be displayed in the terminal.

Important Notes:

• Closing the terminal or losing the client connection will terminate the test process.

• To run the test as a background daemon, execute:

    > ./rep-benchmark.sh > /dev/null 2>&1 &

• To monitor progress, check the logs:

    > cd ./logs
    > tail -f log_info.log

4. Use Case

We take the application of CRRC Qingdao Sifang Vehicle Research Institute Co., Ltd. as an example, and refer to the scene described in "Apache IoTDB in Intelligent Operation and Maintenance Platform Storage" for practical operation instructions.

Test objective: Simulate the actual needs of switching time series databases in the scene of CRRC Qingdao Sifang Institute, and compare the performance of the expected IoTDB and KairosDB used by the original system.

Test environment: In order to ensure that the impact of other irrelevant services and processes on database performance and the mutual influence between different databases are eliminated during the experiment, the local databases in this experiment are deployed and run on multiple independent virtual servers with the same resource configuration. Therefore, this experiment set up 4 Linux (CentOS7 /x86) virtual machines, and deployed IoT-benchmark, IoTDB database, KairosDB database, and MySQL database on them respectively. The specific resource configuration of each virtual machine is shown in Table 4-1. The specific usage of each virtual machine is shown in Table 4-2.

Table 4-1 Virtual machine configuration information

Hardware Configuration Information	Value
OS system	CentOS7
number of CPU cores	16
memory	32G
hard disk	200G
network	Gigabit

Table 4-2 Virtual machine usage

IP	Usage
172.21.4.2	IoT-benchmark
172.21.4.3	Apache-iotdb
172.21.4.4	KaiosDB
172.21.4.5	MySQL

4.1 Write Test

Scenario description: Create 100 clients to simulate 100 trains, each train has 3000 sensors, the data type is DOUBLE, the data time interval is 500ms (2Hz), and they are sent sequentially. Referring to the above requirements, we need to modify the IoT-benchmark configuration parameters as listed in Table 4-3.

Table 4-3 Configuration parameter information

Parameter Name	IoTDB Value	KairosDB Value
DB_SWITCH	IoTDB-013-SESSION_BY_TABLET	KairosDB
HOST	172.21.4.3	172.21.4.4
PORT	6667	8080
BENCHMARK_WORK_MODE	testWithDefaultPath
OPERATION_PROPORTION	1:0:0:0:0:0:0:0:0:0:0
CLIENT_NUMBER	100
GROUP_NUMBER	10
DEVICE_NUMBER	100
SENSOR_NUMBER	3000
INSERT_DATATYPE_PROPORTION	0:0:0:0:1:0
POINT_STEP	500
OP_MIN_INTERVAL	0
IS_OUT_OF_ORDER	false
BATCH_SIZE_PER_WRITE	1
LOOP	10000
TEST_DATA_PERSISTENCE	MySQL
TEST_DATA_STORE_IP	172.21.4.5
TEST_DATA_STORE_PORT	3306
TEST_DATA_STORE_DB	demo
TEST_DATA_STORE_USER	root
TEST_DATA_STORE_PW	admin
REMARK	demo

First, start the tested time series databases Apache-IoTDB and KairosDB on 172.21.4.3 and 172.21.4.4 respectively, and then start server resource monitoring through the ser-benchamrk.sh script on 172.21.4.2, 172.21.4.3 and 172.21.4.4 (Figure 4-1). Then modify the conf/config.properties files in the iotdb-0.13-0.0.1 and kairosdb-0.0.1 folders in 172.21.4.2 according to Table 4-3 to meet the test requirements. Use benchmark.sh to start the writing test of Apache-IoTDB and KairosDB successively.

Figure 4-1 Server monitoring tasks

For example, if we first start the test on KairosDB, IoT-benchmark will create a CONFIG data table in the MySQL database to store the configuration information of this test (Figure 4-2), and there will be a log output of the current test progress during the test execution (Figure 4-3) . When the test is completed, the test result will be output (Figure 4-3), and the result will be written into the FINAL_RESULT data table (Figure 4-4).

Figure 4-2 Test configuration information table

Figure 4-3 Test progress and results

Figure 4-4 Test result table

Afterwards, we will start the test on Apache-IoTDB. The same IoT-benchmark will write the test configuration information in the MySQL database CONFIG data table. During the test execution, there will be a log to output the current test progress. When the test is completed, the test result will be output, and the result will be written into the FINAL_RESULT data table.

According to the test result information, we know that under the same configuration the write delay times of Apache-IoTDB and KairosDB are 55.98ms and 1324.45ms respectively; the write throughputs are 5,125,600.86 points/second and 224,819.01 points/second respectively; the tests were executed respectively 585.30 seconds and 11777.99 seconds. And KairosDB has a write failure. After investigation, it is found that the data disk usage has reached 100%, and there is no disk space to continue receiving data. However, Apache-IoTDB has no write failure, and the disk space occupied after all data is written is only 4.7G (as shown in Figure 4-5); Apache-IoTDB is better than KairosDB in terms of write throughput and disk occupation. Of course, there will be other tests in the follow-up to observe and compare from various aspects, such as query performance, file compression ratio, data security, etc.

Figure 4-5 Disk usage

So what is the resource usage of each server during the test? What is the specific performance of each write operation? At this time, we can visualize the data in the server monitoring table and test process recording table by installing and using Tableau. The use of Tableau will not be introduced in this article. After connecting to the data table for test data persistence, the specific results are as follows (taking Apache-IoTDB as an example):

Figure 4-6 Visualization of testing process in Tableau

4.2 Query Test

Scenario description: In the writing test scenario, 10 clients are simulated to perform all types of query tasks on the data stored in the time series database Apache-IoTDB. The configuration is as follows.

Table 4-4 Configuration parameter information

Parameter Name	Example
CLIENT_NUMBER	10
QUERY_DEVICE_NUM	2
QUERY_SENSOR_NUM	2
QUERY_AGGREGATE_FUN	count
STEP_SIZE	1
QUERY_INTERVAL	250000
QUERY_LOWER_VALUE	-5
GROUP_BY_TIME_UNIT	20000
LOOP	30
OPERATION_PROPORTION	0:1:1:1:1:1:1:1:1:1:1

Results:

Figure 4-7 Query test results

4.3 Description of Other Parameters

In the previous chapters, the write performance comparison between Apache-IoTDB and KairosDB was performed, but if the user wants to perform a simulated real write rate test, how to configure it? How to control if the test time is too long? Are there any regularities in the generated simulated data? If the IoT-Benchmark server configuration is low, can multiple machines be used to simulate pressure output?

Table 4-5 Configuration parameter information

Scenario	Parameter	Value	Notes
Simulate real write rate	OP_INTERVAL	-1	You can also enter an integer to control the operation interval.
Specify test duration (1 hour)	TEST_MAX_TIME	3600000	The unit is ms; the LOOP execution time needs to be greater than this value.
Define the law of simulated data: support all data types, and the number is evenly classified; support five data distributions, and the number is evenly distributed; the length of the string is 10; the number of decimal places is 2.	INSERT_DATATYPE_PROPORTION	1:1:1:1:1:1	Data type distribution proportion
LINE_RATIO	1	linear
SIN_RATIO	1	Fourier function
SQUARE_RATIO	1	Square wave
RANDOM_RATIO	1	Random number
CONSTANT_RATIO	1	Constant
STRING_LENGTH	10	String length
DOUBLE_LENGTH	2	Decimal places
Three machines simulate data writing of 300 devices	BENCHMARK_CLUSTER	true	Enable multi-benchmark mode
BENCHMARK_INDEX	0, 1, 3	Take the writing parameters in the write test as an example: No. 0 is responsible for writing data of device numbers 0-99; No. 1 is responsible for writing data of device numbers 100-199; No. 2 is responsible for writing data of device numbers 200-299.