This preprint proposes a Memory DBMS: a database management system for long-term institutional memory shared by humans, applications, and AI agents. It addresses the problem that organizations store records, documents, tickets, logs, and messages, but often lose the evidence, timing, permissions, reasoning, and workflow context needed to reuse that knowledge safely. The paper defines a memory-native architecture for storing Memory Objects with source links, provenance, temporal validity, permissions, audit records, confidence, precision, lifecycle state, and workflow relationships. It presents the object model, retrieval model, security and governance requirements, implementation requirements, and evaluation plan for a scalable Memory DBMS.
Honour et al. (Sun,) studied this question.