Introduction to ClickHouse: Features, Installation, Performance Testing, and Comparison

1. Introduction

ClickHouse is an open‑source column‑oriented database developed by Yandex for real‑time analytical workloads, offering 100‑1000× faster processing than traditional row‑based systems.

2. Key Features

Fast: utilizes all hardware, peak query performance >2 TB/s (decompressed).

Fault‑tolerant: asynchronous multi‑host replication, no single point of failure.

Scalable: vertical and horizontal scaling to thousands of nodes.

Easy to use: SQL‑based, no custom APIs required.

Hardware‑efficient: columnar storage reduces I/O by up to 10‑100×.

CPU‑efficient: vectorized execution with SIMD and JIT.

Optimized disk access: minimizes range scans.

Minimizes data transfer: works without specialized high‑performance networks.

3. When to Use / Not Use

Suitable for immutable, well‑structured event or log streams; not suitable for OLTP, high‑value key‑value, blob, or document workloads.

4. Why It Is Fast

OLAP characteristics—read‑heavy, wide tables, batch writes, and columnar storage—allow selective column reads, high compression, vectorized query execution, and reduced disk I/O, all contributing to superior performance.

5. Installation

sudo yum install yum-utils
sudo rpm --import https://repo.clickhouse.tech/CLICKHOUSE-KEY.GPG
sudo yum-config-manager --add-repo https://repo.clickhouse.tech/rpm/stable/x86_64
sudo yum install clickhouse-server clickhouse-client

Configuration directories after yum installation:

/etc/clickhouse-server/   (config files)
/var/lib/clickhouse/        (data files)
/var/log/clickhouse-server/ (log files)

These paths can be edited in /etc/clickhouse-server/config.xml and /etc/clickhouse-server/users.xml .

Start and connect:

sudo /etc/init.d/clickhouse-server start
clickhouse-client -m   # default login as user 'default'
clickhouse-client --user=xxx --password=xxx --host=xxx

6. Performance Test

SELECT C_CITY, S_CITY, toYear(LO_ORDERDATE) AS year, sum(LO_REVENUE) AS revenue
FROM lineorder_flat
WHERE (C_CITY='UNITED KI1' OR C_CITY='UNITED KI5')
  AND (S_CITY='UNITED KI1' OR S_CITY='UNITED KI5')
  AND year>=1992 AND year<=1997
GROUP BY C_CITY, S_CITY, year
ORDER BY year ASC, revenue DESC;

The query processed 546.67 million rows in 1.723 seconds, achieving ~4.46 GB/s throughput (≈317 million rows/s).

7. Comparison with Other Column Stores

Different storage types suit different scenarios; there is no universal “silver bullet.”

8. Adoption

Major internet companies using ClickHouse include:

Toutiao – thousands of nodes, dozens of PB daily.

Tencent – game data analytics with a dedicated ops system.

Ctrip – >80% of business runs on ClickHouse, billions of rows daily.

Kuaishou – ~10 PB total, 200 TB daily, 90% queries < 3 s.

Yandex, CloudFlare, Spotify, and others worldwide.

9. Conclusion

The article provides a concise overview of ClickHouse’s capabilities, installation steps, and performance characteristics, encouraging readers to consult the official documentation for deeper exploration.