Questions: Bloom Filters in Distributed Systems

5 questions to test your understanding

Score: 0 / 5
Question 1 Multiple Choice

A Bloom filter is queried for key X and returns 'NOT IN SET.' What can you conclude with certainty?

AKey X is probably not in the set, but there is a small chance it is (false negative)
BKey X is definitely not in the set — this answer is guaranteed to be correct
CKey X is definitely not in the set on this node, but may exist on other replicas
DNothing certain — the result depends on how many hash functions were used
Question 2 Multiple Choice

A distributed database uses Bloom filters to coordinate anti-entropy between replicas. Node A sends its Bloom filter to Node B. Node B queries the filter for 10,000 keys it holds and finds that 150 are reported as 'IN SET' on Node A. How should Node B interpret this result?

ANode B should send all 10,000 keys to Node A because Bloom filter 'IN SET' answers are unreliable
BNode B should send the remaining ~9,850 keys (those reported 'NOT IN SET' on A) because these are definitely missing from A; the 150 'IN SET' keys are probably present on A but may include some false positives
CNode B should request Node A send its full key list, because the Bloom filter cannot identify which specific keys are missing
DNode B should ignore the result — Bloom filters are only useful for caching, not for replication protocols
Question 3 True / False

A Bloom filter can definitively confirm that an element IS in the set — if most k hash positions return 1, the element was definitely added.

TTrue
FFalse
Question 4 True / False

You can delete an element from a standard Bloom filter by setting its k hash-position bits back to 0, since those bits were originally set during insertion.

TTrue
FFalse
Question 5 Short Answer

Why can't you delete elements from a standard Bloom filter, and what variant addresses this limitation? What tradeoff does the variant introduce?

Think about your answer, then reveal below.