FOUNDATIONAL ABSTRACTIONS FOR CORE ENTITIES AND QUERY MECHANISMS IN EVENT-SOURCED SYSTEMS

Abstract

Event-sourced systems represent application state as a deterministic function of an immutable event history, so queries operate over histories and derived views rather than over a single current-state snapshot. The absence of shared, implementation-agnostic definitions and semantic contracts makes nominally similar queries non-comparable and hinders portable cost reasoning. The goal of the study is to develop a compact, implementation-agnostic formal foundation for core entities and query mechanisms in event-sourced systems, which enables mechanism-level cost analysis independent of specific technologies. A theoretical methodology based on formalization and mechanism-level analysis is proposed. The study defines a minimal set of core entity abstractions that determine representation and interpretation in event-sourced systems (events, streams, aggregates, projections, snapshots, versions). On this basis, querying is formalized as contract-defined deterministic evaluation over immutable histories and organized into four mechanism groups: reconstruction, temporal, cross-stream, and retroactive replay, each specified through explicit scope and cut rules with declared ordering/merge, correlation, and version-normalization policies. Portable cost envelopes are derived by expressing selection and evaluation costs through selected evidence size and amortized per-event processing, including explicit contributions from normalization and replay shortening by snapshots. The study formalized implementation-agnostic core abstractions and contract-defined query mechanisms for event-sourced systems and derived cost envelopes. A theoretical experiment on a synthesized banking event dataset confirmed internal consistency, replay equivalence, and reproducibility under semantics-preserving transformations. The proposed formalization fixes the semantic degrees of freedom required for reproducible and comparable querying over immutable event histories and provides a reusable basis for mechanism-level cost reasoning across implementations. Further research should extend the framework toward practice-complete semantics by formalizing admissibility rules under distributed time uncertainty or read-model staleness under eventual consistency.