solidDB is a relational "SQL Database" with persistent in-memory - and on-disk tables. It provides three different replication features: a highly automated HotStandby (HSB), asynchronous log (CRep) replication with arbitrary topologies (with/wo HSB) and a scalable cluster called the Grid. solidDB can be configured as a separate client/server, used as a library which is less need with the client, or shared as a library with multiple client processes and accessed through shared memory. Recent additions include JSON/BSON support, and vector datatype and vector indexing (with hnswlib).
Conflict Resolution Synchronous Replication Asynchronous Replication
HotStandby can be configured to run with fine-grained levels ranging from asynchronous to strictly synchronous replication.
Asynchronous (CRep) replication detects schema inconsistencies and conflicting updates.
B+Tree Patricia/Radix Trie Hierarchical Navigable Small World (HNSW)
B+-tree for on-disk tables, and combination of compressed Patricia trie and cache conscious B-tree leaf level nodes. Hnswlib uses hardware optimized distance functions if the hardware supports them.
Read Committed Serializable Repeatable Read
Read Committed, Repeatable Read & Serializable for on-disk tables, Read Committed and Repeatable Read for in-memory tables.
Disk-oriented In-Memory Hybrid
Per table granularity. Both in-memory - and on-disk tables are supported on same schema. Table types are transparent to queries.
https://teamblue.unicomsi.com/products/soliddb/
https://support.unicomsi.com/manuals/soliddb/100/index.html
Solid Information Technology
1992
IBM, Unicom Global