The implementation of ClickHouse on the Yandex.Metrica service marks the following indicators: up to 1 billion rows per second on a single server and up to 2 TB per second on a cluster of 400 nodes. Such performance is associated with columnar data storage, physical sorting of data by primary key, parallelization of operations on several processor cores of one server, and support for approximate calculations.
Cross Data Center Replication is possible in ClickHouse, which replicates data between clusters, providing failover protection and high-performance data access for globally distributed and mission-critical Big Data applications.
ClickHouse supports a structured query language dialect close to the ANSI SQL standard, with extensions including arrays and nested data structures, URI function, and the ability to connect external key-value storage.
High availability and fault tolerance
Due to decentralisation and the absence of a single point of failure, ClickHouse works reliably in a distributed cluster. The DBMS maintains the complete identity of data on different replicas, automatically recovering them after failures.
Ease of deployment and convenience of use
As a means of connecting to the DBMS, you can use the console, HTTP API, JDBC and ODBC drivers, as well as many “wrappers” in Python, PHP, NodeJS, Perl, Ruby, and R.
What is Google BigQuery?
Google BigQuery is a PaaS service that supports most of the features of a DBMS.Learn more
What is Amazon Redshift?
Amazon Redshift is columnar cloud-based data warehouse platform, which is useful for systems with a constant load – such as regular reporting and machine learning routines.Learn more