Layer-4 Load Balancer Request Flow Quiz

Q1. At a Layer 4 load balancer, which information is used as the key when creating (or looking up) a connection-tracking entry for a new TCP flow?




Q2. In a DNAT/SNAT (a.k.a. one-arm NAT) forwarding mode, what does the load balancer do to the first packet before sending it to the chosen backend?




Q3. Which event usually triggers the load balancer to select a backend and create the conntrack entry for a TCP connection?




Q4. Which Layer 4 forwarding mode lets the backend server send its responses **directly** to the client without passing back through the load balancer?




Q5. When a Layer 4 load balancer operates as a full TCP proxy, how many distinct TCP connections exist for a single client flow?




Q6. For a pure UDP service handled by an L4 LB, what typically determines how long a flow’s conntrack entry is kept before being removed?




Q7. Which statement best describes the amount of application-layer inspection a classic Layer 4 load balancer performs?




Q8. Why does a Layer 4 load balancer run active health checks against its backend targets?




Q9. Consistent-hash algorithms like Google-Maglev are often used in L4 load balancers because they primarily provide which benefit?




Q10. Enabling source-IP stickiness on a Layer 4 load balancer causes which change in behaviour?




system-design