codejump

Distributed Systems — Series 3

Preview — 3 of 10 questions

A client needs 20 independent small records. Each network round trip has high fixed overhead. The API can accept a bounded list of IDs. Which change best reduces that overhead?

AIncrease each response's compression level.
BCache all responses permanently after the first fetch.
CFetch them sequentially with a longer timeout.
DFetch the IDs in one bounded batch request.

A batch read requests 50 independent resources; two are missing. Clients need the remaining 48 results. Which API contract is most useful?

AReturn per-item outcomes with stable item identifiers.
BReject the batch whenever one requested item is missing.
CReturn a success status and silently omit missing IDs.
DRetry the entire batch until all 50 resources exist.

A report must traverse exactly the dataset as it existed when export began, despite concurrent inserts and deletes. What is required beyond a stable cursor ordering?

AA secondary index on the display label only.
BA retained snapshot or equivalent versioned read view.
CAn opaque cursor that encodes only the last row's ID.
DA larger page size on the live dataset.

Sign up free to play

Answer all 10 questions (7 more), see explanations for every answer, and track your score.