Chapter 26 — Data governance: architecture alternatives
Lesson map
Lesson 26 published. Map, example, EX076.
flowchart TD
N0["Learning outcomes"]
N1["26.1 Three options"]
N2["26.2 Evaluation"]
N3["26.3 Costs and skills"]
N4["26.4 Recording the choice"]
N0 --> N1 --> N2 --> N3 --> N4
N4 --> EX["Worked example"]
EX --> Q["Exercise EX076"]
Logical drawing
flowchart LR
N0["Learning outcomes"]
N1["26.1 Three options"]
N2["26.2 Evaluation"]
N3["26.3 Costs and skills"]
N4["26.4 Recording the choice"]
N0 --> N1 --> N2 --> N3 --> N4 --> Q["EX076"]
[Learning outcomes]
|
v
[26.1 Three options]
|
v
[26.2 Evaluation]
|
v
[26.3 Costs and skills]
|
v
[26.4 Recording the choice]
|
v
[Worked example] --> [Exercise EX076]
Learning outcomes
You will (1) compare architectures for governed data; (2) evaluate them against requirements; (3) record the choice.
26.1 Three options
Kofi compares three architectures for Tidewater's governed data. A central warehouse with a governance layer holds all curated data in one place. A data mesh gives each domain team responsibility for publishing its data as a product with agreed standards. A hybrid keeps a central warehouse for regulatory reporting while domains own operational data. He describes each option's data flows, ownership model and the controls each would need to satisfy the requirements from chapter 21. Each option is drawn with the same notation for comparison.
26.2 Evaluation
He scores each option against the ranked requirements. The central warehouse scores well on consistency and reconciliation but poorly on ownership, because the central team becomes de facto owner. The mesh scores well on ownership but poorly on reconciliation across domains at Tidewater's current maturity. The hybrid scores best overall: regulatory figures stay reconciled centrally, while domain teams own and publish their source data with standard quality rules and metadata. Kofi shows the scoring sheet to owners so they can challenge any score they consider unfair.
26.3 Costs and skills
Kofi estimates costs and skills. The mesh requires every domain team to have data engineering skills, which Tidewater lacks outside the analytics team. The hybrid requires training for domain owners and stewards but not new engineering teams. He includes a three-year plan for building skills, after which more responsibility could move to domains. He notes that architecture choices must fit the organisation's capabilities, not just its aspirations, or they fail during implementation. The plan names training courses, mentors and checkpoints at which progress will be reviewed formally.
26.4 Recording the choice
Kofi writes an architecture decision record choosing the hybrid, listing the alternatives, scores, costs, risks and conditions for revisiting the decision, such as domain teams reaching agreed skill levels. The governance board approves it. Prof. Varga notes that recording conditions for revisiting a choice keeps architecture decisions alive, since they encourage the organisation to reconsider when circumstances change rather than defending an outdated choice out of habit. The record is stored with the other decision records, and the board diary schedules a review after eighteen months to check whether the conditions have been met.
Worked example
Kofi compares central, mesh and hybrid architectures, scores them against requirements, chooses the hybrid given Tidewater's skills and records conditions for revisiting the decision.
Synthetic case. Options are synthetic.
Errors and limits
Fashionable architectures may not fit; score against your requirements.
Exercises
- EX076 — Mesh weakness. Why did the mesh score poorly at Tidewater? Rubric: 3 = weak cross-domain reconciliation at current maturity;; 2 = partly correct;; 1 = vague;; 0 = none.
- EX077 — Chosen option. Which architecture was chosen? Rubric: 3 = the hybrid;; 2 = partly correct;; 1 = other;; 0 = none.
- EX078 — Revisit condition. Give one condition for revisiting the decision. Rubric: 3 = domain teams reaching agreed skill levels;; 2 = partly correct;; 1 = vague;; 0 = none.
Continue studying — lesson 26 (chapter 26 of 50)
← Chapter 25 · Course chapter index · Chapter 27 →
Continue to the next authored lesson.