An OLAP engineered for web scale
ClickHouse was developed internally by Yandex (Russian search engine) from 2009 to power Yandex.Metrica, a web analytics platform of Google Analytics-like scale. Released as open source in June 2016 under Apache 2.0 licence. In September 2021 ClickHouse Inc. is founded (spin-off with ex-Yandex leadership and VC investments), US-based.
Technical features
- Columnar storage — data organised by column instead of row, ideal for analytical queries touching few fields across many rows
- MergeTree engine — background compaction of sorted segments, like LSM-tree adapted for OLAP
- Compression — LZ4, ZSTD, Delta, Gorilla — reduces storage 5-10x vs. raw data
- SIMD vectorisation — batch processing with modern CPU instructions
- Materialised views with incremental aggregations
- Distributed — sharded + replicated clusters with ZooKeeper/ClickHouse Keeper
- Standard SQL — supports much ANSI SQL with analytical extensions
Performance
Public benchmarks show ClickHouse handling queries on billions of rows in a few seconds on medium clusters. Competes directly with Apache Druid, Apache Pinot, Amazon Redshift, Google BigQuery, Snowflake on analytical scenarios.
Usage
- Web analytics (Yandex, Cloudflare, Uber scale)
- Observability — log aggregation, APM (hyperscaler backend)
- Ad-tech — real-time bidding analytics
- Financial — price time series
- IoT — sensor data ingest and query
- E-commerce — behavioural analytics
In the Italian context
Growing Italian adoption in:
- Ad-tech — Italian advertising platforms
- Gaming studios
- Telco — CDR/NetFlow analysis
- Italian SaaS analytics
References: ClickHouse open source (June 2016), Yandex. ClickHouse Inc. (September 2021). Apache 2.0 licence. MergeTree engine. Columnar storage. Yandex.Metrica as first user. Competitors: Druid, Pinot, Redshift, BigQuery, Snowflake.