ByteDance Java Backend Internship Interview Questions and Answers

Recent students often ask how prepared they need to be for a large‑tech daily internship, and the answer is that the interview covers the same core topics as campus recruitment: fundamentals, projects, and algorithms.

The backend section emphasizes programming language proficiency, MySQL, Redis, network protocols, operating system knowledge, backend project experience, and algorithmic thinking, while noting that advanced topics like message queues, distributed systems, and micro‑services can be deferred.

Network

What is the difference between HTTP and HTTPS?

HTTP transmits data in plaintext; HTTPS adds SSL/TLS encryption between TCP and HTTP.

HTTPS requires an additional TLS handshake after the TCP three‑way handshake.

Default ports differ: 80 for HTTP, 443 for HTTPS.

HTTPS needs a digital certificate from a Certificate Authority.

How do domain names map to IP addresses?

Through DNS resolution: the client queries a local DNS server, which may consult root, top‑level, and authoritative servers to finally obtain the IP address.

What is the difference between GET and POST?

GET retrieves resources and places parameters in the URL (ASCII only, length limited), while POST sends data in the request body without size restrictions.

Explain TCP three‑way handshake

Client in CLOSE state sends SYN (client_isn) to server.

Server in LISTEN state replies with SYN‑ACK (server_isn, ack=client_isn+1).

Client acknowledges with ACK (ack=server_isn+1) and both sides enter ESTABLISHED.

The three‑handshake prevents old duplicate connections, synchronizes initial sequence numbers, and avoids resource waste.

Operating System

How to kill a process by name?

ps -ef | grep <process_name>   // find PID<br/>kill <PID>                     // terminate process

What IPC mechanisms exist?

Anonymous and named pipes (unidirectional, FIFO).

Message queues (structured messages, kernel‑linked).

Shared memory (fastest, requires synchronization via semaphores).

Signals (asynchronous notifications).

Sockets (TCP/UDP or local communication).

What does kill -9 use?

It sends the SIGKILL signal.

How does fork work?

Creates a child process with a copy‑on‑write page table.

Parent receives child's PID; child receives 0.

Both share physical memory until a write triggers a copy.

Explain Copy‑On‑Write

Parent and child share pages marked read‑only; on write, the kernel copies the page, allowing both processes to modify their own copies without blocking each other.

Java

ArrayList vs LinkedList

ArrayList uses a dynamic array; O(1) random access.

LinkedList uses a doubly‑linked list; O(n) random access.

Insert/delete at ends is O(1) for ArrayList, O(1) for LinkedList at any position, but LinkedList uses more memory.

When to use each?

Use ArrayList for frequent random reads; use LinkedList for frequent insertions/deletions.

HashMap internal structure

Pre‑Java 8: array + linked list; after 8 entries, converts to a red‑black tree. Java 8+ uses array + linked list/red‑black tree.

ConcurrentHashMap thread safety

Java 7 uses segment locks (array of Segment objects); Java 8 uses lock‑striped bins with volatile, CAS, and synchronized for updates, reducing contention.

Redis

Common data structures

String (SDS), List (quicklist), Hash, Set, Sorted Set (skiplist).

Why use Redis instead of local cache?

Higher query efficiency, reduces MySQL load.

Rich data types, clustering, distributed locks.

Handling memory limits

Configure maxmemory and choose an eviction policy (noeviction, volatile‑*, allkeys‑*).

Expiration strategies

Redis combines lazy deletion (via expireIfNeeded) and periodic active expiration (randomly checks 20 keys per cycle, stopping when expired‑key ratio < 25%).

RDB persistence

RDB creates a full snapshot using bgsave (forked child, copy‑on‑write) so the main thread continues handling commands; save blocks the main thread.

MySQL

CHAR vs VARCHAR

CHAR: fixed length, always occupies declared space.

VARCHAR: variable length, stores only actual data plus length bytes.

Index classifications

By data structure: B‑tree, hash, full‑text.

By storage: clustered (primary) vs secondary.

By field property: primary, unique, normal, prefix.

By column count: single‑column vs composite.

Index optimization tips

Use prefix indexes for long strings.

Prefer covering indexes to avoid table lookups.

Make primary keys auto‑increment to reduce page splits.

Avoid situations that cause index invalidation.

EXPLAIN output fields

possible_keys

, key, key_len, rows, type (All → const).

Watch extra for Using filesort, Using temporary, Using index.

Algorithm

Tree level order traversal.

Maximum width of a binary tree.

Merging k sorted arrays.