Assessment of Change Costs

✅

general aspects are discussed broadly and so well known. we need to consider and evaluate thos for all out parts of kube3

Core idea: Changes usually disrupt. Evaluate a tool before commitment — not just on features, but on how hard it is to leave or change. Change is usually only welcome if promis of improving. 2026-05-22--13-57-41

for general hints

Cloud Risk Assessment — Dimensions

Dimension	What to ask
Permanence / Constancy	Is the service likely to exist and behave the same in 5 years?
Product governance	Is direction externally controlled, or is there a realistic Right to Fork? Does the vendor add features without asking?
Persistence	Are identifiers durable? (URLs, DOIs, domains, GUIDs, hashes — only sufficiently long random strings qualify)
Transferability / Portability	Open standards? (OCM, IMAP, ODT, …) Operable elsewhere? Switchable to another product? api-first, conformance tests like at SCS
Subsidiarity	No forced centralisation without good reason

Applied to this project

Variant A — Syncthing

Dimension		Guessing
Permanence	seems ok ✅	Mature, stable protocol; slow-moving by design
Governance	seems ok ✅	Open source (MPL-2); realistic fork; no vendor adding features unilaterally
Persistence	seems ok ✅	No external identifiers; file paths and device IDs under your control
Portability	seems ok ✅	Runs on several os includign windows and hardware; open protocol; no cloud dependency
Subsidiarity	✅	Fully decentralised by design; no central server required

Assessment: low change cost — Syncthing data and config are fully portable.

Variant B — Nextcloud

Dimension		Guessing
Permanence	✨⚠️/ 🟨 🟩	Fast-moving product; frequent breaking changes natural and occured
Governance		AGPL, fork possible in principle; opencloud
Persistence	❌	Share-links are tied to domain; no stable identifiers without custom setup (portable DNS helps)
Portability	⚠️	Container migration feasible but domain-bound; federation via OCM partially helps
Subsidiarity		Central server required; no local-only mode. but anyone can run an own server ✅

Assessment: moderate–high change cost — domain dependency is the main risk for link persistence and migration.

Variant C — Syncthing + Nextcloud (combined)

Dimension		Guessing
Permanence	✅	Syncthing anchors data persistence; Nextcloud layer is replaceable
Governance	✅	Syncthing shields data from Nextcloud volatility
Persistence	⚠️	File content: ✅ via Syncthing. Share-links: ❌ still domain-bound via Nextcloud
Portability		Data survives Nextcloud instance loss; restore = redeploy container + Syncthing sync
Subsidiarity	✅	Syncthing layer is fully decentralised; Nextcloud is opt-in on top /below it

Assessment: low-to-moderate change cost — Syncthing absorbs most portability concerns; main residual risk is Nextcloud share-link continuity. This applies only to the common things like files.

← Contact · Milestones →

tfc-i1q · Export created: 2026-07-28 12:16 UTC