Java Backend Interview Guide: Networks, Data Structures, MySQL Logs, Elasticsearch, Kafka, Microservices, and Spring Essentials

Interview Topics Overview

The guide summarizes key interview questions for a Java backend position, emphasizing the need to master each technology stack rather than just superficial usage.

Network

DNS operates over UDP because it offers low latency, simplicity, and lightweight headers, with reliability handled by retries and caching.

HTTP characteristics: text‑based, extensible, flexible, and stateless, requiring mechanisms like cookies or sessions to maintain state.

Data Structures & Algorithms

Array vs. Linked List: access O(1) vs. O(n), insertion/deletion O(1) vs. O(n), cache locality, and appropriate use cases.

Implementing a queue with two stacks (stackPush and stackPop) – push to stackPush, pop/peek by transferring elements to stackPop when needed. The amortized complexity is O(1) for all operations.

import java.util.Stack;

class MyQueue {
    private Stack<Integer> stackPush;
    private Stack<Integer> stackPop;

    public MyQueue() {
        stackPush = new Stack<>();
        stackPop = new Stack<>();
    }

    public void push(int x) { stackPush.push(x); }

    public int pop() {
        if (stackPop.isEmpty()) {
            while (!stackPush.isEmpty()) {
                stackPop.push(stackPush.pop());
            }
        }
        return stackPop.pop();
    }

    public int peek() {
        if (stackPop.isEmpty()) {
            while (!stackPush.isEmpty()) {
                stackPop.push(stackPush.pop());
            }
        }
        return stackPop.peek();
    }

    public boolean empty() { return stackPush.isEmpty() && stackPop.isEmpty(); }
}

// Test code
public class Main {
    public static void main(String[] args) {
        MyQueue queue = new MyQueue();
        queue.push(1);
        queue.push(2);
        System.out.println(queue.peek()); // 1
        System.out.println(queue.pop());  // 1
        System.out.println(queue.empty()); // false
    }
}

Backend Components

MySQL logs: redo log (transaction durability), binlog (replication & backup), undo log (rollback).

Elasticsearch full‑text search relies on inverted indexes that map terms to document IDs.

Kafka guarantees no data loss via producer acknowledgments, replicated partitions, and consumer offset management; it avoids duplicate consumption through idempotent producers, unique keys, or downstream deduplication (e.g., Redis or DB unique constraints).

Microservices

Common components: Eureka (service registry), Ribbon (client‑side load balancing), Feign (declarative HTTP client), Hystrix (circuit breaker), Zuul (API gateway), Config (configuration center).

Load‑balancing algorithms: simple round‑robin, weighted round‑robin, random, weighted random, consistent hashing, and least‑active.

Service circuit breaking and degradation concepts explained with Hystrix and fallback strategies.

Spring Essentials

IoC (Inversion of Control) – container creates, initializes, and destroys beans, decoupling object creation from business logic.

Why field injection is discouraged: external visibility, potential circular dependencies, and inability to inject final or static fields.

AOP fundamentals: aspects, join points, advice types, pointcuts, weaving, and proxy mechanisms (JDK dynamic proxy vs. CGLIB).

Transactional proxy behavior – self‑invocation via this bypasses the proxy, so transactions are not applied.

Spring resolves singleton setter‑based circular dependencies using a three‑level cache (singleton factories, early references, fully initialized beans).

Spring MVC request flow: DispatcherServlet → HandlerMapping → HandlerAdapter → Controller → ModelAndView → ViewResolver → View rendering → response.