Foundations & Layering of Load Balancing Quiz

Q1. Users in different continents are experiencing high latency when accessing a service hosted in a single data center. How can load balancing help reduce response times for these users? Install a Layer 2 switch in the network to optimize packet routing for long-distance connections.
Deploy servers in multiple regions and use a global load-balancing strategy (e.g., geoDNS or anycast) to route each user's request to the nearest server location.
Increase the number of servers in the original data center so each user's request is handled faster by the expanded pool.
Have users adjust their network OSI-layer settings on their devices to better accommodate long-distance traffic.

Q2. After implementing a new load balancer, one backend server is handling significantly more traffic than the others even though all servers have identical capacity. Which configuration issue is MOST likely the cause? Health-check intervals are too frequent, overwhelming one server with probes instead of distributing them.
The virtual IP (VIP) was not configured on the load balancer, so only one server is receiving traffic.
The load balancer has session persistence (sticky sessions) enabled, causing many returning users to stick to the same server.
The load balancer is using a simple round-robin algorithm to distribute traffic evenly among servers.

Q3. An organization wants users worldwide to connect to a single URL and be automatically directed to the nearest data center for lower latency. Which load-balancing technique best accomplishes this? Deploying a Layer 7 load balancer in one central data center and routing all traffic through it.
Using an Anycast IP address announced from multiple data centers so that user traffic is routed to the nearest location.
Requiring users to manually select their geographic region on the website before connecting.
Configuring a single large server to proxy and forward requests to all other data centers as needed.

Q4. Which scenario is an example of Layer 7 (application-layer) load balancing? Using BGP anycast routing to send users to the nearest server location without regard to HTTP details.
Distributing traffic using round-robin DNS entries without inspecting any request content.
Routing HTTP requests to different server pools based on the URL path in the request.
Forwarding network packets to servers based only on destination IP address and port number.

Q5. A new web service uses HTTP/3, which runs over QUIC (UDP). The existing L7 proxy supports only HTTP/1.1 and HTTP/2 over TCP. What must be true for the load balancer to handle HTTP/3 traffic? No changes are needed; if the load balancer supports HTTP/2, it will automatically support HTTP/3.
The service must force HTTP/3 clients to downgrade to HTTP/1.1 or HTTP/2 so the load balancer can process the traffic.
Replace the load balancer with a DNS-based strategy, since DNS is required for any QUIC traffic.
The load balancer must be capable of handling UDP at the transport layer (either via an upgrade or a new load balancer that supports QUIC).

Q6. After switching to a Layer 7 load balancer, backend logs show the load balancer's IP as the client IP. Why, and how can this be fixed? The L7 load balancer terminates the client connection and opens a new one to the server, so the server sees the LB's IP. Enable X-Forwarded-For headers or Proxy Protocol to pass along the original client IP.
The load balancer is using NAT at Layer 4, so the original IP is lost; switching to Layer 7 mode would resolve it.
It is a fundamental requirement of the OSI model that the client IP be hidden when passing through a load balancer.
This wouldn't actually happen; load balancers always preserve the original client IP by default.

Q7. A team wants to offload SSL/TLS encryption work from backend servers by terminating TLS at the load balancer. What kind of load balancer do they need? A transport-layer (Layer 4) load balancer, because it can pass through encrypted traffic without handling certificates.
An application-layer (Layer 7) load balancer that can terminate SSL/TLS and inspect HTTP requests before forwarding.
A DNS load balancer, since DNS strategies distribute TLS handshakes across servers.
Any load balancer will automatically handle TLS offloading by default.

Q8. A Layer 4 load balancer routes traffic by forwarding packets to backend servers without modifying application content. What information does it primarily use to make routing decisions? HTTP headers such as Host or URL path.
Network addresses and transport ports (e.g., client IP, destination IP, and TCP/UDP port numbers).
Application-layer data like specific strings in the message body.
The contents of encrypted payloads after decrypting TLS traffic.

Q9. During a sudden backend server failure, the load balancer stops sending traffic to that server. What mechanism enables this? The TCP protocol automatically reroutes traffic when one server in the pool does not respond.
A built-in firewall blocks traffic to the failed server's IP.
Regular health checks that probe backend servers and mark a failed server as out of rotation.
The client's browser detects the server is unresponsive and notifies the load balancer.

Q10. A low-latency financial service is concerned that adding a load balancer might increase response times. Which type generally introduces the least latency overhead? A DNS-based load-balancing approach, since a DNS lookup occurs for each user request.
An application-layer proxy performing deep content inspection and data transformations.
A Layer 7 load balancer that fully parses and analyzes each HTTP request.
A Layer 4 load balancer that simply forwards packets between client and server.

Q11. To achieve zero-downtime deployment, which strategy best uses the load balancer? Deploying without a load balancer, as it cannot help with deployment processes.
Gradually draining traffic from servers being updated by marking them out of service on the load balancer.
Reducing DNS TTL to zero right before deployment so clients rapidly switch to new servers.
Restarting all servers simultaneously because the load balancer will queue requests.

Q12. A stateful web application needs all requests from a specific user to go to the same backend server during their session. Which load-balancing feature addresses this? Anycast routing, because the same user will always connect to the geographically nearest data center.
Session persistence (sticky sessions) using cookies or client IP.
DNS load balancing, since DNS will always direct a user to the same server once resolved.
A basic Layer 4 load balancer with a purely random selection algorithm.

Q13. Your microservice uses a custom UDP-based protocol. Which type of load balancer is appropriate? An HTTP(S) reverse proxy, because HTTP protocols are similar enough to handle UDP traffic.
A DNS round-robin load balancer, because UDP traffic should be balanced via DNS.
A Layer 4 load balancer, as it operates at the transport layer and can handle UDP traffic.
A Layer 7 load balancer, since it can inspect arbitrary UDP payloads in detail.

Q14. Which of the following correctly describes the flow of a client request through a typical load-balanced web service? The client rotates its requests through each backend server without contacting a load balancer.
The client opens connections to all backend servers simultaneously, and the server that responds fastest handles the request.
The load balancer sends the client a list of backend IPs, and the client directly chooses which server to contact.
The client resolves the service's domain to a VIP of a load balancer, connects to the load balancer, which forwards the request to a backend server and returns the response.

Q15. A company uses DNS load balancing to direct clients to the nearest regional data center, then Layer 7 load balancing within each region. Which benefit does this multi-tier architecture provide? It eliminates the need for health checks at the load balancer level.
It reduces user latency and isolates failures to a region by keeping traffic local to the nearest servers.
It simplifies DNS management because records never need to change.
It guarantees no single backend server will ever become overloaded.

Q16. A critical service uses a pair of load balancer instances sharing a Virtual IP (VIP); the standby takes over if the active fails. What does this achieve? Ensures the service remains reachable at the same IP (high availability) even if one load balancer fails.
Balances load simultaneously at both the network (L3) and application (L7) layers.
Improves performance by splitting the VIP into two physical IPs.
Eliminates the need for health checks on backend servers.

Q17. During a sale event, an e-commerce site adds a load balancer in front of multiple identical servers. Which primary goal of load balancing does this address? Reducing development time by requiring fewer servers.
Minimizing network latency by shortening physical network paths.
Distributing traffic across multiple servers (horizontal scaling) to handle more load and avoid a single point of failure.
Encrypting and decrypting all traffic to secure communications.

Q18. DNS-based load balancing is configured across two regions. When one region fails, some clients still try to connect there for several minutes. Why? DNS caching at clients or ISPs causes them to keep using the last resolved IP until the DNS TTL expires.
The working region's load balancer is not performing health checks on servers in the failed region.
The load balancer was configured at the wrong OSI layer, preventing region failover.
Clients use an outdated OSI model that doesn't support multi-region failover.

system-design