Best Columnar Databases for Sematext Cloud

The Apache Cassandra database provides high availability and scalability without compromising performance. It is the ideal platform for mission-critical data because it offers linear scalability and demonstrated fault-tolerance with commodity hardware and cloud infrastructure. Cassandra's ability to replicate across multiple datacenters is first-in-class. This provides lower latency for your users, and the peace-of-mind that you can withstand regional outages.

ClickHouse is an open-source OLAP database management software that is fast and easy to use. It is column-oriented, and can generate real-time analytical reports by using SQL queries. ClickHouse's performance is superior to comparable column-oriented database management software currently on the market. It processes hundreds of millions of rows to more than a million and tens if not thousands of gigabytes per second. ClickHouse makes use of all hardware available to process every query as quickly as possible. Peak processing speed for a single query is more than 2 Terabytes per Second (after decompression, only utilized columns). To reduce latency, reads in distributed setups are automatically balanced between healthy replicas. ClickHouse supports multimaster asynchronous replication, and can be deployed across multiple datacenters. Each node is equal, which prevents single points of failure.

MariaDB Platform is an enterprise-level open-source database solution. It supports transactional, analytical, and hybrid workloads, as well as relational and JSON data models. It can scale from standalone databases to data warehouses to fully distributed SQL, which can execute millions of transactions per second and perform interactive, ad-hoc analytics on billions upon billions of rows. MariaDB can be deployed on prem-on commodity hardware. It is also available on all major public cloud providers and MariaDB SkySQL, a fully managed cloud database. MariaDB.com provides more information.

Apache HBase™, is used when you need random, real-time read/write access for your Big Data. This project aims to host very large tables, billions of rows and X million columns, on top of clusters of commodity hardware.