**Why TCP in Data Pipelines:** Reliable, ordered delivery—critical for DB connections, Kafka (uses TCP), file transfer. Guarantees: no loss, no duplicates, ordered.
**Mechanics:** 3-way handshake (SYN, SYN-ACK, ACK); flow control (sliding window); congestion control; retransmission on loss. Overhead: ~40B header, RTT for handshake.
**Trade-offs:** UDP = lower latency, no guarantee—use for metrics, streaming where loss acceptable. TCP = reliability....
The complete answer continues with detailed implementation patterns, architectural trade-offs, and production-grade considerations. It covers performance optimization strategies, common pitfalls to avoid, and real-world examples from companies like ZS Associates. The answer also includes follow-up discussion points that interviewers commonly explore.
Continue Reading the Full Answer
Unlock the complete expert answer with code examples, trade-offs, and pro tips - plus 1,863+ more.
Or upgrade to Platform Pro - $39
Engineers who used these answers got offers at
AmazonDatabricksSnowflakeGoogleMeta
According to DataEngPrep.tech, this is one of the most frequently asked Python/Coding interview questions, reported at 1 company. DataEngPrep.tech maintains a curated database of 1,863+ real data engineering interview questions across 7 categories, verified by industry professionals.