System Design test (Advanced)

Sample questions

5 sample questions from the System Design test, with the answers. Samples are taken out of the live test, so you will not meet these when you sit it.

1. A company is building an e-commerce platform and needs to ensure that the order processing system can handle high traffic without impacting the inventory management system. Which of the following is the most appropriate approach?

   1. Implement a relational database to store order and inventory data in a single schema.
   2. Store all order and inventory data in a distributed file system for faster access.
   3. Use a REST API with synchronous calls for all communication between the systems.
   4. Use a message queue broker to decouple the order processing system from the inventory management system.

   Answer: D. Use a message queue broker to decouple the order processing system from the inventory management system.

   A message queue broker allows systems to communicate asynchronously, which is essential in scenarios like order processing and inventory management. This decoupling ensures that order processing does not block inventory management and vice versa. The other options either do not leverage message queues or are not directly related to the use of message queue brokers.

2. Which statement accurately describes a common mistake or tricky edge case in implementing distributed rate limiting?

   1. Distributed rate limiting can be implemented using a token bucket algorithm with a shared, centralized counter.
   2. Distributed rate limiting is often implemented using a centralized counter with a lock, which can become a bottleneck as traffic increases.
   3. Distributed rate limiting is best achieved by storing rate limit data in a distributed cache with no consistency guarantees.
   4. Distributed rate limiting can be implemented using a single node with a local cache, which is efficient and scalable.

   Answer: B. Distributed rate limiting is often implemented using a centralized counter with a lock, which can become a bottleneck as traffic increases.

   Distributed rate limiting requires a mechanism that can scale across multiple nodes without a single point of failure. Using a centralized counter with a lock is a common mistake because it introduces a bottleneck and single point of failure, making it unsuitable for high-throughput distributed systems. A better approach is to use a distributed data store or consensus algorithm to manage the rate limits across nodes.

3. Which caching strategy is best suited for a system that requires horizontal scaling and low-latency access to frequently used data across multiple nodes?

   1. Distributed caching layers
   2. Read-through caching
   3. Cache eviction policies
   4. Write-through caching

   Answer: A. Distributed caching layers

   Distributed caching layers are designed to scale horizontally and reduce latency by storing copies of data across multiple nodes. This ensures that data can be accessed from the nearest node, improving performance. The other options do not directly address the horizontal scaling and latency reduction benefits of distributed caching.

4. You are designing a distributed system where multiple services need to access a shared dataset with high read frequency and occasional updates. What approach should you use to ensure cache coherence across all services?

   1. Use a distributed cache with a write-through strategy and cache invalidation on updates.
   2. Use a cache-aside strategy with no TTL on cache entries.
   3. Use a write-behind strategy without cache invalidation.
   4. Use a read-through strategy with no eviction policy.

   Answer: A. Use a distributed cache with a write-through strategy and cache invalidation on updates.

   A write-through strategy ensures that the cache and the underlying database are always in sync, while cache invalidation on updates ensures that stale data is removed from the cache whenever the data is modified. This approach helps maintain coherence across distributed systems by preventing inconsistencies between the cache and the source of truth (the database). Other options either ignore cache coherence or do not address the issue effectively.

5. What is canary deployment?

   1. Canary deployment rolls out changes to all users at once to ensure maximum coverage.
   2. Canary deployment gradually rolls out changes to a subset of users or servers to test performance and stability before a full rollout.
   3. Canary deployment is a strategy where changes are rolled back immediately if any issue is detected.
   4. Canary deployment is used exclusively for database migrations and not for application updates.

   Answer: B. Canary deployment gradually rolls out changes to a subset of users or servers to test performance and stability before a full rollout.

   Canary deployment is a strategy where a new version of an application is rolled out to a small subset of users or servers to test its performance and stability before a full deployment. This allows teams to detect issues early and roll back if needed, minimizing the risk of widespread failures. Other options may describe different deployment strategies like blue-green deployment or rolling updates, but they do not align with the definition of canary deployment.