Question 1

What are the key differences between HTTP/1.1, HTTP/2, and HTTP/3?

Accepted Answer

HTTP/2 adds multiplexing and header compression over TCP; HTTP/3 uses QUIC (UDP-based) to eliminate head-of-line blocking and improve connection establishment. HTTP/1.1: one request per TCP connection (or pipelining, poorly supported). HTTP/2: multiplexes streams over one TCP connection, server push, HPACK header compression. HTTP/3: QUIC over UDP eliminates TCP HOL blocking — one lost packet doesn't block other streams.

Question 2

What is head-of-line (HOL) blocking and which HTTP version solves it at the transport layer?

Accepted Answer

HOL blocking occurs when a lost TCP packet stalls all multiplexed HTTP/2 streams — HTTP/3 solves this at the transport layer by using QUIC (UDP) where each stream is independent. HTTP/2 multiplexes over one TCP connection. But TCP is reliable — a lost packet stalls the entire connection until retransmitted, blocking all streams. QUIC maintains per-stream flow control: a lost packet only stalls its stream, not others.

Question 3

What is consistent hashing and why is it used in distributed systems?

Accepted Answer

A hashing scheme where adding/removing a node only remaps ~1/n of keys to adjacent nodes — minimizing cache invalidation and data movement compared to modulo hashing. Modulo hashing: key % N — adding a node remaps most keys. Consistent hashing: nodes and keys are placed on a ring; a key maps to the next clockwise node. Adding a node only steals keys from its successor. Used in Cassandra, DynamoDB, CDNs, and Memcached.

Question 4

What is the TCP TIME_WAIT state and why does it matter for high-volume servers?

Accepted Answer

After actively closing a connection, the socket stays in TIME_WAIT for 2×MSL (~60s) to absorb delayed packets from the old connection — can exhaust ports on high-connection-rate servers. A server making millions of short-lived outgoing connections (e.g., to a database) can exhaust all 64K ports if they're all in TIME_WAIT. Mitigations: SO_REUSEADDR, net.ipv4.tcp_tw_reuse=1, connection pooling, or using persistent connections.

Question 5

What is a reverse proxy and what benefits does it provide?

Accepted Answer

A proxy server sitting in front of backend servers — handling SSL termination, load balancing, caching, compression, and hiding server topology from clients. nginx as reverse proxy: terminates TLS (offloading expensive crypto from app servers), load balances across instances, caches static content, rate limits, and hides internal IPs. Clients only ever see the reverse proxy's address.

Computer Networks — Senior Dev Interview Questions

More Senior Dev Topics

Computer Networks at Other Levels