ClickHouse on Hikube

Hikube ClickHouse databases offer an open-source, high-performance, column-oriented SQL management system, designed for online analytical processing (OLAP). They guarantee rapid ingestion of massive data, execution of complex queries in near real-time, and the reliability necessary for critical enterprise analytical applications.

---

🏗️ Architecture and Operation

ClickHouse architecture is based on two essential parameters that allow adapting deployment to real needs:

• Shards → they allow distributing data into multiple pieces across different nodes. The more shards there are, the more the load is distributed, which improves query execution speed on very large volumes.
• Replicas → they create redundant copies of shards. This increases resilience and fault tolerance, while allowing read load distribution across multiple nodes.

🔎 Illustrative Example

Imagine a database of 1 billion customer records:

• 1 shard – 1 replica
  All data is stored in a single space.
  Use cases:

  • Proof of concept (POC)
  • Development environments
  • Occasional analytical workloads

• 2 shards – 1 replica
  Data is divided into two parts (e.g., customers A–M and N–Z). Queries are executed in parallel, which significantly speeds up analysis.
  Use cases:

  • Analysis on large data volumes
  • Applications requiring better performance
  • Regular reports on large customer bases or transactions

• 2 shards – 2 replicas
  Each shard is duplicated on another node. You benefit from both speed (distributed data) and security (fault tolerance).
  Use cases:

  • Critical analytical applications in production
  • High availability needs
  • Multi-user platforms with high query concurrency
  • Disaster recovery plans (DRP)