Storage

Apache Iceberg in S3

Table data in Tengri is stored in the S3 cluster in the open format Apache Iceberg.

  • Iceberg supports INSERT, UPDATE, and DELETE operations, as well as table schema changes.

  • Iceberg supports ACID, a set of requirements for a transactional system to ensure that it operates in the most reliable and predictable manner.

  • Iceberg-structure is formed based on Parquet format with column-by-column data compression.

  • Iceberg is thousands of successful implementations by the world’s largest companies to work with hundreds of TB and dozens of PB data.

Columnar storage format

By utilising the Parquet format, Tengri supports:

Columnar storage format

OLAP- queries run faster by reading only the required table columns.

Column-by-column data compression

OLAP- queries work faster by a factor of 5 or more by compressing table data.