FUTURECENTRAL PRESS · STUDY GUIDE SAMPLE
Blockchain-based Supply Chains in Agribusiness
A Manager’s Guide to Traceability, System Design, and Governance
Chapter 7: Choosing the Architecture and Governance Model
How to read a chain’s structure into a choice of network and governance, add tamper-evidence at a price a cooperative can sustain, and use a database where a database suffices
Study Guide Sample
The companion Study Guide supports revision and exam preparation. Each chapter guide includes a summary, ten MCQs with an answer key, five short-answer questions with model answers, and three essay questions with structured hints.
This sample contains the chapter summary, four selected MCQs with answers, three short-answer questions with model answers, and two essay questions with hints.
Chapter summary
Architecture is read off the chain, never picked from a menu. Three facts settle it. Which parties must write decides permissioned or public, and in agri chains those writers are a small, identified set of businesses, so the answer comes out permissioned nearly every time. Who is already trusted decides how much work the ledger does and how far control spreads. Who should govern decides where value lands and who bears the cost.
The Architecture-and-Governance Choice takes permissioned as settled and offers three models, which Chapter 10 resolves into six. A single-owner corporate anchor governs the network, ties its own hands so partners get commitment, and keeps the efficiency. A cooperative puts control, and value with it, into producer hands. The third, a neutral utility, runs as shared infrastructure no participant controls, letting competitors join. The model is really a decision about value capture in other clothes. The test of whether any is real is the node topology and the change-control process: a cooperative where one member quietly operates all the nodes is a single-owner network under a label.
The Minimal-Ledger Pattern makes the chosen network affordable. The conventional stack keeps working, and hashes are anchored only where a record has to be tamper-evident across parties, so the ledger holds fingerprints and stays cheap. Anchoring one fingerprint per batch costs a fraction of settling millions of entries on chain. For a federation with a thin surplus, that decides whether cross-party tamper-evidence can be paid for.
Right-sizing places each record on a rung. A database serves wherever a trusted party holds records the others accept. The Minimal-Ledger Pattern fits wherever parties genuinely need tamper-evidence. A single-owner ledger earns its keep wherever the point is credible commitment, since a database its owner can silently edit gives partners no assurance.
Multiple-choice questions
1. Four competing seed companies need a shared certification registry that all of them will write to and rely on. Which governance model fits?
Answer key: Question 1
Answer: C
2. A network's charter reserves upgrade approval to a two-thirds member vote, but the upgrade key is held by one member's technology arm, which also operates every node. What is the network?
Answer key: Question 2
Answer: A
3. A federation records several million milk entries a day. Why does anchoring hashes cost so much less than settling those entries on chain?
Answer key: Question 3
Answer: C
4. A cautious board will approve one of two first-year plans. Plan One anchors the member-payment record and leaves everything else on the existing systems. Plan Two migrates procurement, payment, and quality records onto the ledger together. Which does the minimal-ledger pattern support?
Answer key: Question 4
Answer: D
Short-answer questions and model answers
1. A team presents this design for a producer federation: a permissioned network, cooperative governance, member payment records anchored as hashes, and the conventional payment system left in place. Name three things the design leaves unspecified, and say what each omission exposes.
Read the model answer
Model answer. The design is silent on the node and change-control arrangement. Cooperative governance is a claim about who holds control, and control lives in who operates the nodes and who can deploy an upgrade, so the federation is exposed to a network that is cooperative on the charter and single-owner in operation. It is silent on first-mile integrity. What the milk analyzer or the scale reports at the collection center is the fact every payment rests on, and anchoring its hash secures the record without making the reading true. It is silent on where member personal data sits. Placement is a legal constraint set before building, so a design that has not stated it can drift into anchoring member identities on an immutable record, which is a privacy breach and a design error no later correction undoes.
2. Give the strongest case a well-informed skeptic could make against the minimal-ledger pattern, and then say where that case fails.
Read the model answer
Model answer. The skeptic's case is that the pattern buys almost nothing. Anchoring a hash proves only that a record matches what was written at anchoring time, and leaves open whether the record was right or whether anyone will ever check it. The bottom rung concedes the point: across a great many agri record-keeping tasks, a well-run database in the hands of a trusted party is the right answer, and these members have accepted the federation's database for years. On that reading the anchor is a cost with no reader behind it. The case is strong on what anchoring does not do and weak on the one thing it does. The database rung holds only while every reader accepts the keeper's authority. The moment a member wants a check the federation cannot influence, the database offers nothing, provided the anchor's nodes and change control sit outside the federation's sole hands.
3. A federation moves its member-payment record from the database rung to the minimal-ledger rung. Trace what changes for the federation's finance team, for a member farmer, and for an external buyer that has asked for independent assurance.
Read the model answer
Model answer. For the finance team, almost nothing changes operationally. The payment system still computes and pays, and the added work is anchoring a fingerprint per batch and holding records so they can be re-hashed later. For the member, a disputed entry can be re-hashed against an anchor the federation cannot alter on its own, so long as nodes and change control sit outside its sole hands. That turns a complaint into a matter of arithmetic. For the buyer, the record turns from assertion into evidence. It can test any claim the federation makes about a member payment without commissioning an audit, which is assurance the federation's own database could never supply. That is the move from the database rung to the minimal-ledger rung, and it is the buyer's demand that pays for it.
Long-answer questions and essay hints
1. A poultry integrator has built this architecture. A permissioned network under single-owner governance, with the integrator operating all four nodes. Full flock, feed, and veterinary records written on chain. Contract growers are admitted as read-only participants, grower payment is computed off-chain by the integrator's own system with the result anchored daily, and grower identity, bank details, and farm location are written on chain so that any participant can verify a consignment's origin.
Attack this design. Identify the three points at which it is most likely to fail, rank them by likelihood, and say what evidence would confirm each. (600 to 900 words)
Read the essay hints
- Frameworks to deploy: the Architecture-and-Governance Choice and its control test, the right-sizing rungs, and the Minimal-Ledger Pattern.
- Work through the design element by element before you rank anything, since the ranking has to rest on a reading you have already made.
- Rank the three by likelihood and defend the ordering, which matters more than the list itself.
- For each, state the evidence that would confirm it is happening, expressed as something an auditor could look for.
- Finish by saying which single element you would change first if the integrator would grant you only one.
- Word count: 600 to 900 words.
- Pitfall to avoid: treating single-owner governance as the flaw. That model is defensible where an anchor's commitment is the point, so the argument has to turn on what this design does with it.
2. A cotton chain runs from about 12,000 growers through four ginners to three exporters, and from there on to a set of overseas brands that buy the finished yarn. The growers trust the ginners to weigh honestly and distrust them on grade. The exporters are trusted by the brands for solvency and by nobody for origin claims. A state marketing board is trusted by the growers and regarded by the exporters as slow and politically exposed.
Choose a governance model for the shared record. Devote most of your answer to justifying the two models you reject. (600 to 900 words)
Read the essay hints
- Frameworks to deploy: the Architecture-and-Governance Choice across all three models, read against the three facts about the chain.
- Lead with the trust distribution as the chain actually presents it, including the places where trust is split by function.
- Reject the two models you will not use, and make each rejection specific to this chain, since a model that fails elsewhere may fit here.
- Name what your chosen model still fails to solve, since the split trust here means no model fits cleanly.
- End on the governance detail that would have to carry what the model cannot.
- Word count: 600 to 900 words.
- Pitfall to avoid: rejecting a model on a weakness this chain does not have. Each rejection should turn on a fact given in the scenario.