Reliable UDP

What this reliable-UDP design question tests

This is an easy-to-medium networking design question that probes whether you understand the core gaps between UDP and TCP, and how to bridge them pragmatically. It's common in interviews for systems roles at trading firms and infrastructure teams, where ultra-low-latency communication matters but reliability cannot be sacrificed.

The question asks you to think architecturally rather than implement code. A strong answer identifies the specific failure modes UDP introduces—packet loss, reordering, duplication—and describes concrete mechanisms to detect and recover from each. Interviewers listen for whether you understand trade-offs: adding reliability layers costs latency and complexity, so you should articulate which tools (sequence numbers, acknowledgements, timeouts, checksums) you'd actually deploy and why.

• Packet loss detection and retransmission
• Sequence numbering and reordering
• Acknowledgement schemes and flow control
• Timeout and retry strategies
• Checksum and integrity validation