Branimir Lambov

Log-structured merge-tree databases rely on a combination of in-memory buffers and hierarchies of immutable on-disk structures, which are two cases where a trie data structure is most efficient.

Choosing a compaction strategy for a Cassandra database has historically been a very difficult problem, where making the wrong choice can have lasting effects on performance while making a change later is a time-consuming and costly process. The Unified Compaction Strategy, introduced with Cassandra 5, is designed to provide a solution to this problem by effectively handling a diverse range of use

Memtables have a very profound effect on the performance of Cassandra, as writes to the memtable are on the hot write path, and its large data structures remain on the Java heap for a long time and are a key determinant of garbage collection performance.

Cassandra 5 will include a new compaction strategy which aims to provide a common solution for the kinds of workloads that the existing compaction strategies in Cassandra target, solve the existing strategies’ deficiencies, improve performance and permit easy reconfiguration. The new unified strateg

Byte order, the property of being comparable using lexicographic comparison on the bytes of a key, and tries, structures implementing ordered maps using byte-labelled transitions, have been in use in DataStax Enterprise since version 6 and are currently being contributed to mainline Cassandra in CEP