RWS Architecture article

ADR-009 ODF Structure for 1H2026 Engagement

The long-term storage design assumes ODF Advanced External, but the 1H2026 engagement timeline might favor a simpler interim deployment. We need to choose the structure that meets

  1. Typeadr
  2. Statusaccepted
  3. Domainstorage
On this page
  1. Context and Problem Statement
  2. Decision Drivers
  3. Considered Options
  4. Decision Outcome
  5. Pros and Cons of the Options
  6. Option 1 – ODF hyperconverged
  7. Option 2 – ODF Advanced External (chosen)

Context and Problem Statement

The long-term storage design assumes ODF Advanced External, but the 1H2026 engagement timeline might favor a simpler interim deployment. We need to choose the structure that meets delivery timelines without blocking migration to the long-term model.

Decision Drivers

  • Alignment with the accepted storage backend in ADR-001.
  • Time-to-deliver for the 1H2026 engagement.
  • Ability to migrate to the long-term external ODF design.

Considered Options

  1. ODF hyperconverged deployment.
  2. ODF Advanced External (dedicated Ceph platform).

Decision Outcome

Chosen option: ODF Advanced External (dedicated Ceph platform).

Implementation baseline:

  • Production storage integration uses ODF external mode backed by dedicated Ceph clusters.
  • Hyperconverged ODF is not a production target and is allowed only for lab/prototyping scenarios.
  • Migration and interoperability checks between ODF and external Ceph are mandatory release gates for storage changes.

Pros and Cons of the Options

Option 1 – ODF hyperconverged

Pros

  • Faster initial deployment with fewer external dependencies.

Cons

  • Per-cluster storage duplication and harder migration to externalized storage.

Option 2 – ODF Advanced External (chosen)

Pros

  • Aligns with the dedicated Ceph platform strategy and shared storage model.

Cons

  • Requires dedicated storage cluster readiness and integration work.