Routing and load balancing in Microservices - Practice Problems & Coding Challenges

Challenge - 5 Problems

🧠 Conceptual

intermediate

How does round-robin load balancing distribute requests?

Imagine you have 3 identical servers behind a load balancer using round-robin. How will the load balancer distribute 6 incoming requests?

A Requests are sent to the server with the least CPU usage.

B All 6 requests go to Server1 until it is full, then move to Server2.

C Requests are sent randomly to any server without order.

D Each server receives 2 requests in order: Server1, Server2, Server3, then repeat.

💡 Hint

Think about how round-robin cycles through servers evenly.

Round-robin load balancing sends each request to the next server in a fixed order, cycling through all servers evenly.

❓ Architecture

intermediate

Choosing a load balancer for microservices

You have a microservices system with many small services. Which load balancing approach fits best to route requests efficiently?

A Use DNS round-robin to distribute requests among service IPs.

B Use client-side load balancing where each service instance chooses the next service instance to call.

C Use a centralized hardware load balancer that routes all requests externally.

D Use a single service instance to route all requests to others.

💡 Hint

Think about scalability and dynamic service discovery in microservices.

Client-side load balancing lets each service instance decide where to send requests, improving scalability and handling dynamic changes.

❓ scaling

advanced

Scaling load balancers for high traffic

Your system faces millions of requests per minute. What is the best way to scale load balancing to avoid bottlenecks?

A Use a single powerful load balancer with very high CPU and memory.

B Use only client-side load balancing without any central load balancer.

C Deploy multiple load balancers in parallel with DNS-based load balancing among them.

D Route all traffic through a database to decide which server to use.

💡 Hint

Think about avoiding single points of failure and distributing load evenly.

Multiple load balancers in parallel distribute traffic and avoid bottlenecks or single points of failure.

❓ tradeoff

advanced

Tradeoffs between sticky sessions and stateless load balancing

What is a key tradeoff when using sticky sessions (session affinity) in load balancing compared to stateless balancing?

A Sticky sessions simplify user session handling but reduce load distribution flexibility.

B Stateless balancing causes user sessions to be stored on the load balancer.

C Stateless balancing requires all requests to go to the same server.

D Sticky sessions improve fault tolerance but reduce cache efficiency.

💡 Hint

Consider how session data affects routing choices.

Sticky sessions keep a user tied to one server, simplifying session state but limiting even load distribution and failover.

❓ estimation

expert

Estimating load balancer capacity for peak traffic

Your service expects 10 million requests per hour at peak. Each request requires 5ms processing on the load balancer. How many load balancer instances are needed to handle peak load without queuing?

A At least 14 load balancer instances

B At least 28 load balancer instances

C At least 7 load balancer instances

D At least 35 load balancer instances

💡 Hint

Calculate requests per second and how many can be handled per instance.

10 million requests/hour = ~2778 requests/second. Each instance handles 1/0.005 = 200 requests/second. 2778/200 ≈ 13.9, so 14 instances needed.