UCloud Kuaijie UDB vs Self‑Built MySQL: Performance & Cost Comparison

Introduction

The author’s infrastructure uses a mix of UCloud RDS and self‑hosted MySQL databases. With the launch of UCloud’s Kuaijie‑based UDB, which claims significant CPU, OS kernel, disk, and network improvements, a comparative performance test was conducted to guide future instance selection.

What Is Kuaijie?

Kuaijie is a cloud host that adds a distributed storage layer on top of physical machines, optimizes the OS kernel, and leverages high‑performance CPUs and networking. Its key specifications include:

• Flagship cloud host with excellent compute, storage, and network performance
• Supports AMD EPYC 2.9 GHz or Intel Cascadelake 2.5 GHz CPUs, up to 96 cores and 768 GB RAM
• Network throughput up to 1000 W PPS, storage IOPS up to 120 W
• Flexible instance sizes, suitable for web services, gaming, databases, and data‑analysis workloads

Disk space can be expanded to 32 TB, addressing common storage‑shortage issues during database operations.

Environment Preparation

Both test targets run MySQL 5.7.25. The RDS instance uses the default my.cnf template with the following core parameters (unchanged from UCloud defaults):

| binlog_format                | ROW |
| innodb_adaptive_flushing    | ON  |
| innodb_adaptive_hash_index  | ON  |
| innodb_buffer_pool_instances| 8   |
| innodb_buffer_pool_size     | 8589934592 |
| innodb_file_per_table       | ON  |
| innodb_flush_log_at_trx_commit | 2 |
| innodb_flush_method         | O_DIRECT |
| innodb_io_capacity          | 15000 |
| innodb_lock_wait_timeout    | 50 |
| innodb_log_buffer_size      | 8388608 |
| innodb_log_file_size        | 1073741824 |
| innodb_log_files_in_group   | 2 |
| innodb_max_dirty_pages_pct | 50.0 |
| innodb_open_files           | 1024 |
| innodb_read_io_threads      | 8 |
| innodb_write_io_threads     | 8 |
| innodb_stats_on_metadata   | OFF |
| innodb_thread_concurrency   | 20 |
| max_binlog_cache_size       | 18446744073709547520 |
| max_binlog_size             | 1073741824 |
| max_connect_errors          | 1000000 |
| max_connections             | 2000 |
| max_user_connections        | 0 |
| open_files_limit            | 1000000 |
| sync_binlog                 | 1 |
| table_definition_cache      | 464 |
| table_open_cache            | 128 |
| thread_cache_size           | 50 |

The test machines were equipped with 10 Gbps network interfaces, Intel CPUs (details shown in the images), and measured network latency of 0.15‑0.22 ms to the UDB instance and 0.22‑0.30 ms to the self‑built UHost.

Test Scenario

100 tables, each with 100 k rows (total 1 GB), database memory set to 32 GB.

The sysbench 0.5 script used is:

#!/bin/bash
threads="12
24
36
48
60
72
84
96
108"

dt=`date +"%Y%m%d%H%M%S"`
mkdir -p /root/yace_$dt
/root/doDBA -mysql -log &

for i in $threads; do
    echo $i
    time=`date +"%Y%m%d%H%M%S"`
    /opt/yz-sysbench/bin/sysbench \
        --test=/opt/yz-sysbench/share/sysbench/oltp.lua \
        --oltp-tables-count=100 \
        --oltp-table-size=100000 \
        --mysql-db=sysbench \
        --mysql-user=sysbench \
        --mysql-password=sysbench \
        --mysql-host=xx.xx.xx.xx \
        --mysql-port=3306 \
        --max-time=500 \
        --max-requests=0 \
        --oltp-test-mode=complex \
        --num-threads=$i run > /root/yace_$dt/thread_$i_$time.log
    sleep 60
done
p=`pidof doDBA`
kill -9 $p

Test Results

QPS

Insert/Delete

Update

Overall, the Kuaijie‑based UDB delivers more than 20 % higher QPS than the self‑built MySQL on UHost.

The IO‑focused tests showed larger variance, possibly due to short test duration and cloud‑disk jitter; a firmware upgrade is planned for a follow‑up test.

Price Comparison

UDB high‑availability instance costs 5 160 CNY/month, while two Kuaijie UHost machines cost 5 602 CNY/month. The self‑built option also requires additional operational tools (backup, monitoring, HA), increasing total cost.

Thus, from a price‑performance perspective, the UDB solution is more economical.

Conclusion

Advances in CPU, disk, and network technologies have driven rapid improvements in cloud RDS performance. UCloud’s Kuaijie UDB meets the capacity needs of most workloads (excluding extreme bad‑SQL cases) and offers a noticeable performance uplift and better cost efficiency, making it a strong choice for database users.