Which migration can fit six weeks?
Copying a smaller dataset can bound a first phase. It does not replace a transaction system.
The question
Larkspur has 42 days and a 42-day transaction API freeze. All data sizes, allowances and results are invented. The copy model is constant-rate and excludes other delivery work.
| Value | Unit | |
|---|---|---|
| Full dataset | 6000 | decimal GB |
| Read projection | 600 | decimal GB |
| Transfer allowance | 100 | GB/day |
| Validation after copy | 5 | days |
These are authored planning allowances, not measured provider performance. Integration, mapping and operating acceptance are not included.
“Calculate the two scopes, then state what the smaller scope actually enables.”
Not saved by this site. Your browser may restore text when you navigate back.
A strong answer, step by step
Open each step in turn. This is one strong answer, not the only one.
Step 1: What do these inputs establish?
What would you do here? Open the move you’d make.
Compute 65 and 11 modeled days
The strong move. These include the relevant copy time and the stated validation allowance.
Count only the validation days
Tempting, but it misses something. That omits data transfer.
Change GB/day to GB/hour
Tempting, but it misses something. No higher rate was supplied.
See how a strong candidate says it
“Full copy plus validation is 6000/100 + 5 = 65 days. The projection is 600/100 + 5 = 11 days. These are limited planning allowances.”
Why it works: Data size, rate and post-copy validation must use consistent units.
Step 2: What can the smaller copy accomplish?
What would you do here? Open the move you’d make.
Replace the transactions with the projection
Tempting, but it misses something. A read model is not the transaction system.
Promise all delivery work in 11 days
Tempting, but it misses something. The model contains no allowance for the remaining integration or acceptance work.
Plan a read-only first phase and retain the remaining work
The strong move. This matches the smaller scope to a possible customer capability.
See how a strong candidate says it
“It can support planning a read-only phase, if that is the customer need. A projection does not supply complete transactional behavior, and the copy model is not a full delivery forecast.”
Why it works: The data boundary determines the capability that can be claimed.
Step 3: Operations confirms that status reads are enough at the deadline. What boundary follows?
What would you do here? Open the move you’d make.
Rewrite the frozen transaction API
Tempting, but it misses something. This conflicts with the freeze and adds unneeded first-phase scope.
Keep core writes and validate the projection
The strong move. This serves the clarified read need without claiming a transaction replacement.
Increase the transfer estimate until the full dataset fits
Tempting, but it misses something. A preferred answer cannot supply a new transfer allowance.
See how a strong candidate says it
“Preserve transaction writes in the core. Validate the read projection against its actual freshness and mapping contracts rather than expanding to a frozen API rewrite.”
Why it works: The clarified customer need supports a bounded component strategy.
Step 4: The event feed has an unowned legacy status-code mapping. Can you promise the phase?
What would you do here? Open the move you’d make.
Assign the mapping and contract checks before promising delivery
The strong move. This addresses the named dependency instead of hiding it in the date.
Ship because the copy model leaves spare days
Tempting, but it misses something. The mapping has no supported effort or correctness evidence.
Treat copied rows as equivalent business behavior
Tempting, but it misses something. Data movement does not prove semantic equivalence.
See how a strong candidate says it
“Not yet. Assign the mapping and representative contract checks, obtain an effort estimate, and include operating acceptance. Spare modeled copy days do not fund or validate that work.”
Why it works: A semantic dependency can block a plan that fits the copy calculation.
The answer, in short
- Decision
- Use the projection only as a candidate read phase, with mapping and operating evidence still required.
- Because
- The 65-day full-copy model exceeds 42; the 11-day projection allowance does not establish the rest of delivery.
- What would change it
- A changed deadline, measured transfer plan, supported mapping effort or a different customer behavior can change the scope.
Common traps
- Changing the transfer unit.
- Treating read data as transaction behavior.
- Promising the whole project from a copy-only model.
Check your answer
- Did I keep the units consistent?
- Did I state what the copied scope enables?
- Did I respect the API freeze?
- Did I retain the mapping and operating work?
Last reviewed Nov 10, 2026