Router-pattern host · Migration

XyDromatics VNA Migration — archive-to-archive at scale.

Multi-petabyte archive-to-archive migration. Multi-source ingestion from any DICOM-conformant PACS or VNA, per-study cryptographic verification, scheduled migration windows that respect your clinical operations, in-flight tag transformation for legacy-data cleanup, and the audit trail a regulated migration demands.

Capabilities

Built for multi-petabyte migrations.

A migration platform earns its keep over months, not minutes. Per-study verification, pause/resume, scheduled windows, stakeholder reporting — the capabilities below are the ones that distinguish a real migration tool from a clever cron script.

Multi-source ingestion

Read from any DICOM-conformant source — and from non-DICOM exports too.

  • C-STORE SCU (push from source PACS) + SCP (pull to migration host)
  • Q/R against the source archive (C-FIND + C-MOVE / C-GET)
  • STOW-RS for modern HTTP-based source systems
  • File-watcher ingest for offline / disconnected sources
  • Bulk DICOMDIR + filesystem-walk for raw exports
  • Configurable concurrency + rate limits per source

Migration Engine front-end (recommended pattern)

For migrations >100 TB, deploy Migration Engine as a horizontal pre-processor in front of VNA Migration. Cheap edge work happens at the front-end, freeing VNA Migration to focus on verification and audit.

  • Normalization at the edge — character sets, malformed headers, deprecated tag VR mismatches, non-spec timestamp formats
  • Tag coercion — force common values into expected formats (accession-number padding, date-string normalization, VR coercion) before they reach VNA Migration
  • Horizontal scaling — multiple Migration Engine workers parallelize source-side reads
  • Quarantine fixable issues at front-end so VNA Migration sees clean data
  • Net effect: higher overall throughput by reducing retry / error-handling load on the audit-heavy VNA Migration host

Per-study verification

Cryptographic proof that every migrated study matches the source. The defense against silent data loss in long-running migrations.

  • SHA-256 checksum captured at source-side extract
  • Re-verified at target-side ingest before commit
  • Mismatches route to a quarantine queue for clinical review
  • Full per-study audit record: source UID, target UID, hash pair, timestamp

Pause / resume + scheduling

Multi-month migrations can't run at maximum throughput 24x7. Schedule the work to fit your operational windows.

  • Pause / resume at any point with full state checkpointing
  • Scheduled migration windows (off-peak only, weekends only, etc.)
  • Per-source bandwidth caps and priority queues

In-flight tag transformation

Migrations are the moment to clean up legacy data — wrong-patient corrections, scheme changes, AE-Title remappings, deprecated vendor tags.

  • Tag-level rewrite (set / unset / map values) per study
  • Per-source transformation rules (different legacy systems, different cleanup needs)

Progress reporting

Multi-month migrations with stakeholders need accountable progress visibility — not just a CLI percent counter.

  • Real-time dashboard — studies/hour, bytes/hour, ETA per source
  • Per-source + per-modality + per-date-range breakdown

Architecture

Source · Migration · Target.

A migration is fundamentally a three-actor system: a source archive being read, a target archive being written, and the migration host orchestrating the transfer with verification on every study.

Source archive Legacy PACS / VNA Filesystem export Q/R · C-STORE · STOW-RS

Common use cases

Five patterns that cover most migrations.

Most engagements combine two or three of these — vendor consolidation paired with cloud migration, or legacy retirement followed by a compliance export of the final state.

Pattern 1: Legacy PACS retirement

Flow

  1. Source: legacy PACS via Q/R or filesystem export
  2. Migration plan: per-modality, per-date-range, prioritized by clinical relevance
  3. Run in scheduled off-peak windows so source PACS isn't impacted

System requirements & sizing

Tier Source size Throughput CPU RAM Typical engagement
Small migration < 100 TB source < 5 TB / day 8 vCPU 32 GB Single-PACS retirement at a community hospital. 1-3 month runway.
Medium migration 100 TB – 1 PB 5 – 25 TB / day 16 vCPU 64 GB Multi-site PACS migration or vendor consolidation. 6-12 month runway.
Large migration 1 PB – 10 PB 25 – 100 TB / day 32 vCPU + parallel workers 128 GB Health-system VNA replacement, cloud migration. 12-18 month runway.
Enterprise migration > 10 PB > 100 TB / day Multiple worker nodes, 64+ vCPU each 256+ GB per node IDN-scale consolidation. 18-36 month runway with phased cutovers.

Licensing

Migration Core

The base migration platform — DICOM source/target, per-study verification, basic scheduling.

  • C-STORE / Q/R / STOW-RS source + target
  • Per-study SHA-256 verification
  • Basic scheduling (off-peak windows)

Documentation

Document Title Notes
DOC-2026-059 Hardware & Software Requirements (VNA) Sizing, OS support, source-system compatibility matrix
DOC-2026-141 DICOM Conformance Statement (VNA Migration) Public — full SOP class + transfer-syntax matrix
DOC-2026-065 EULA Annex (VNA) Annex to the Synthology Master EULA (DOC-2026-063)

Frequently asked

How is VNA Migration different from Migration Engine?

Different scale and different shape -- but they're often deployed together. VNA Migration is the multi-petabyte archive-to-archive workhorse: full HA, parallel worker nodes, multi-month migration plans, per-study verification, transformation, and hash-chain audit.

What's the audit trail?

Per-study record: source SOP Instance UID + source-side hash + target SOP Instance UID + target-side hash + timestamp + operator identity + any transformation applied. Hash-chain audit log covers the entire migration from start to finish.

Talk through your migration.

Tell us your source archive, target archive, total volume, and target cutover date. We’ll come back within one business day with a discovery proposal and a realistic-not-marketing forecast for the migration window.