Skip to content

Quiz

Click "Answer" to reveal the correct answer and explanation. Try to answer before revealing.

CAP Theorem

Q1. What does the P in CAP stand for, and why is it non-negotiable in any real distributed system?

  • A. Performance — slow systems are considered unavailable
  • B. Persistence — data must survive node restarts
  • C. Partition Tolerance — the system continues operating despite network partitions
  • D. Predictability — responses must arrive within a bounded time
Answer

C — Partition Tolerance.

Network partitions — where messages between nodes are lost or delayed — happen in any real distributed system due to hardware failures, congestion, or physical distance. Because partitions cannot be prevented, the practical design choice is between C (refuse requests during partition) and A (continue serving, possibly stale data).

Q2. A distributed database refuses all reads during a network partition to avoid returning stale data. Which CAP category does it belong to?

  • A. CA — Consistent and Available
  • B. CP — Consistent and Partition-Tolerant
  • C. AP — Available and Partition-Tolerant
  • D. None — it violates all three properties simultaneously
Answer

B — CP.

By refusing reads during a partition, the system prioritises consistency (no stale data) over availability. Once the partition heals, service resumes. ZooKeeper and etcd are classic CP examples.

Q3. Which of the following systems are typically classified as AP (Available and Partition-Tolerant)?

  • A. Apache ZooKeeper and etcd
  • B. PostgreSQL and MongoDB
  • C. Apache Cassandra and DNS
  • D. Redis and SQLite
Answer

C — Apache Cassandra and DNS.

Both continue serving requests during partitions, accepting that responses may be temporarily stale (eventual consistency). ZooKeeper and etcd are CP; PostgreSQL and MongoDB are CP by default; Redis in single-node mode is neither (it is CA, unrealistic for production distribution).

Q4. The PACELC theorem extends CAP. What does the ELC part address that CAP ignores?

  • A. Even when a partition heals, the system must re-elect a leader before serving
  • B. Even without a partition, the system must choose between Latency and Consistency
  • C. Every lost connection requires a Checkpoint before normal operation resumes
  • D. Eventual consistency is always preferred over strong consistency at low load
Answer

B — Even without a partition: Latency vs Consistency trade-off.

CAP only describes behaviour during a partition. PACELC adds that even in normal operation, systems must trade off between lower latency (serve quickly, risk slightly stale reads) and stronger consistency (wait for replication agreement before responding). Cassandra is PA/EL; MongoDB is PC/EC.

Q5. A payment service must never allow a double charge. A network partition occurs between the payment node and the replication node. Which design is correct?

  • A. Use an AP store — availability ensures the payment always goes through
  • B. Use a CA store — the single-node guarantee prevents concurrency issues
  • C. Use a CP store — refuse the request during the partition rather than risk an inconsistent state
  • D. Use an AP store with last-write-wins — timestamps prevent duplicates
Answer

C — CP store.

For financial transactions, correctness is non-negotiable. A CP store refuses to process the payment during the partition (the customer gets an error, not a double charge). An AP store might process the payment on two nodes simultaneously before the partition heals, creating a duplicate transaction that cannot be safely undone.

Q6. In which scenario is a CA system (Consistent and Available) actually achievable?

  • A. In a three-node cluster with synchronous replication
  • B. In a geographically distributed system with dedicated links
  • C. Only on a single-machine system or tightly controlled LAN with no partitions possible
  • D. Never — CA is theoretically impossible
Answer

C — Single-machine or tightly controlled LAN.

CA only works when partitions genuinely cannot occur — which is only realistic for a single server or a local area network under strict control. In any internet-connected or geographically distributed deployment, P is effectively mandatory, leaving the real choice between C and A.

Q7. In PACELC terms, how would you classify a system that favours low latency during normal operation and availability during a partition?

  • A. PC/EC
  • B. PA/EC
  • C. PC/EL
  • D. PA/EL
Answer

D — PA/EL.

PA = during a Partition, favours Availability. EL = Else (normal operation), favours Latency. Apache Cassandra is the canonical PA/EL system — it always responds quickly and accepts eventual consistency in both scenarios.

Consensus Algorithms

Q8. What does the FLP Impossibility Theorem prove about distributed consensus?

  • A. Consensus is impossible in any system with more than one node
  • B. In a purely asynchronous system with even one faulty process, no algorithm can guarantee consensus in bounded time
  • C. Leader election always terminates within a fixed number of rounds
  • D. Byzantine failures make consensus impossible even with synchronous networks
Answer

B — No bounded-time guarantee in an async system with ≥1 faulty node.

FLP (Fischer, Lynch, Paterson, 1985) proves this fundamental limit. Real algorithms respond by assuming partial synchrony (delays are bounded eventually) or by sacrificing liveness in adversarial conditions. This is why Raft and Paxos use timeouts as a proxy for failure detection.

Q9. In Raft, what event triggers a new leader election?

  • A. The current leader sends a resign message to the cluster
  • B. A majority of followers vote to remove the current leader
  • C. A follower's election timeout expires without receiving a heartbeat from the current leader
  • D. The cluster detects that the leader's log has diverged from followers
Answer

C — Follower election timeout expires.

Each follower maintains a randomised election timeout. If no heartbeat arrives before it expires, the follower assumes the leader has crashed and starts a new election by incrementing the term and requesting votes. Randomised timeouts prevent multiple followers from starting elections simultaneously.

Q10. Which formula guarantees that any read quorum overlaps with any write quorum, ensuring you always read the latest write?

  • A. W > N / 2
  • B. R > N / 2
  • C. W = R = N
  • D. W + R > N
Answer

D — W + R > N.

If write quorum W plus read quorum R exceeds total replicas N, at least one node that acknowledged the write must also participate in any read quorum. This overlap ensures the latest write is always visible in a quorum read, forming the basis of systems like Dynamo and Cassandra's tunable consistency.