Aggressive Data Compression
Vertica organizes values of similar data types contiguously in memory and on disk enabling aggressive and simultaneous data compression and encoding for dramatic footprint reduction and blazingly fast parallel load and query times. Based on data cardinality and workloads, the Database Designer automatically chooses from over a dozen encoding and compression algorithms. Vertica’s unique sorting capability further enhances the compression of the data. As a result, our customers achieve more with less hardware.
The Benefits of Operating on Encoded Data
Vertica operates on encoded data which dramatically improves analytic performance by reducing CPU, memory, and disk I/O at processing time. Traditional RDBMS systems are not able to operate on compressed data. A typical row-oriented database system will often expand to 3-5X its original raw data size once auxiliary physical structures like indexes and materialized views are added. Vertica on the other hand will end up reducing the original data size to 1/5th or 1/10th its original size even with our high-availability redundancy turned on. Row-oriented databases that add columnar storage lack many of the benefits of Vertica’s compression because their smaller disk blocks are still most typically represented as columns and rows, e.g. mixed data types. Vertica reads and writes data to disk in large 1MB size blocks, which when combined with the compression, allow each I/O operation to deliver more – a common occurrence is the delivery of an entire column of data in a single I/O operation.
Compression by Data Type