Data Model Mapping
Map kdb+ tables to VogazDB time-series or columnar models, symbols to tags or partitions, and q functions to UQL queries.
Migration guide
Move from kdb+ to a multi-model infrastructure platform.
Preserve time-series strengths while adding columnar analytics, document metadata, SQL-style UQL, auto-tiering, and accelerator-ready architecture.
This page frames migration as a controlled parallel-run program, not a risky rewrite. The goal is semantic equivalence, replay validation, and operational confidence.
Map
Translate
Dual-write
Validate
Cutover
VogazDB vs kdb+
Speed plus flexibility, analytics, and future hardware paths.
Aspect
VogazDB
kdb+
Design eraModern heterogeneous computeLegacy single-node roots
Data modelsTime-series, columnar, documentPrimarily time-series
Query modelSQL-style UQLq language
OperationsNative auto-tieringManual patterns vary
FPGA-ready / CPU-centricFPGA-readyCPU-centric
Migration strategy
Use measured parallel adoption instead of a forced rewrite.
Query Translation
Replace q workloads with SQL-style UQL while preserving semantics and validating result equivalence.
Parallel Run
Dual-write live or replayed data, then compare latency, results, replay behavior, and operational impact.
Validation
Use historical replay, benchmark scenarios, tail-latency analysis, and business-level output comparison.
Controlled Cutover
Move selected workloads first, then expand coverage once operations, security, and support are stable.
Mapping guide
Translate concepts before translating systems.
Migration is easier when teams map schema, query semantics, retention, and operational patterns separately.
kdb+ concept
VogazDB target
TableTime-series or columnar model
SymbolTag, partition, or indexed attribute
q functionUQL query or service-level API
Historical databaseWarm or cold lifecycle tier
Operational scriptsManaged deployment and monitoring workflows
Migration review
Benchmark before committing to migration.
Use a representative dataset and agreed query semantics to compare speed, behavior, cost, and operating model.