Two-phase commit (and why to avoid it) in Microservices - Practice Problems & Coding Challenges

Challenge - 5 Problems

🧠 Conceptual

intermediate

What is the main purpose of the two-phase commit protocol?

In a distributed system, what is the primary goal of using the two-phase commit protocol?

A To allow nodes to commit independently without waiting for others.

B To replicate data asynchronously across nodes for eventual consistency.

C To ensure all participating nodes agree to commit or abort a transaction, maintaining atomicity.

D To speed up transaction processing by parallelizing commits without coordination.

💡 Hint

Think about what atomicity means in transactions across multiple systems.

Two-phase commit ensures that either all nodes commit a transaction or none do, preventing partial commits.

❓ Architecture

intermediate

Which component coordinates the two-phase commit process?

In a two-phase commit protocol, which component is responsible for coordinating the commit or abort decision among participants?

A The coordinator node that sends prepare and commit/abort messages.

B Each participant independently decides when to commit.

C A load balancer that distributes requests evenly.

D A database replica that applies changes asynchronously.

💡 Hint

Consider who sends the 'prepare' message in the protocol.

The coordinator manages the transaction by asking participants if they can commit and then instructing them to commit or abort.

❓ scaling

advanced

Why does two-phase commit limit scalability in microservices?

Which of the following best explains why two-phase commit can reduce scalability in microservice architectures?

A It replicates data too quickly, overwhelming the network.

B It allows services to commit independently, increasing inconsistency.

C It uses asynchronous messaging that can lose data under load.

D It requires all services to lock resources and wait, causing delays and blocking.

💡 Hint

Think about what happens when multiple services wait for each other.

Two-phase commit forces all services to lock resources until the commit decision, causing blocking and reducing throughput.

❓ tradeoff

advanced

What is a major tradeoff when avoiding two-phase commit in distributed systems?

If a system avoids two-phase commit to improve performance, what is a common tradeoff it must accept?

A Accepting eventual consistency instead of strong atomic consistency.

B Guaranteeing immediate consistency but sacrificing availability.

C Increasing the risk of deadlocks due to resource locking.

D Reducing network latency by skipping message acknowledgments.

💡 Hint

Consider what consistency guarantees are relaxed when not using two-phase commit.

Avoiding two-phase commit often means systems accept eventual consistency to gain better performance and availability.

❓ estimation

expert

Estimate the impact of two-phase commit on transaction latency

Assume a distributed transaction involves 5 microservices communicating over a network with an average round-trip time of 50ms. Using two-phase commit, estimate the minimum additional latency introduced by the protocol before the transaction commits.

A About 50ms because only one message round-trip is needed.

B At least 100ms due to prepare and commit phases involving multiple message exchanges.

C Less than 50ms since messages are sent in parallel.

D No additional latency because commits happen instantly.

💡 Hint

Remember two-phase commit requires a prepare phase and a commit phase with message exchanges.

Two-phase commit involves at least two rounds of communication (prepare and commit) with all participants, adding multiple network delays.