Understanding Synchronous vs Asynchronous, Blocking vs Non-Blocking, and Java I/O Models (BIO, NIO, IO Multiplexing, AIO)

The article introduces the complex topics of synchronous vs. asynchronous communication, blocking vs. non‑blocking execution, and various Java I/O models, aiming to present them in simple, understandable language.

1. Synchronous and Asynchronous

同步 means a request does not return until the result is obtained; the caller waits for completion.

异步 means a request returns immediately with no result; the caller can perform other work and later check the status or receive a callback.

Examples are given using a washing machine analogy.

2. Blocking and Non‑Blocking

阻塞 indicates that the current thread is suspended (blocked) while waiting for a result, doing nothing else.

非阻塞 indicates that the thread continues to run and can handle other tasks while the request is pending.

The article explains how blocking relates to thread states such as BLOCKED and RUNNABLE , and how synchronous communication typically keeps the thread in RUNNABLE until the result arrives, whereas asynchronous communication allows the thread to proceed.

3. I/O Models

Java I/O operations involve two stages: waiting for data to become ready and copying data from the kernel buffer to the process address space. Based on these stages, several I/O models exist:

BIO (Blocking I/O)

NIO (Non‑Blocking I/O)

IO Multiplexing (select, poll, epoll)

AIO (Asynchronous I/O)

3.1 BIO

In BIO, the thread is blocked during both stages of the I/O operation; the request is sent to the kernel and the thread remains suspended until the kernel returns data.

3.2 NIO

NIO introduces a non‑blocking first stage where the thread repeatedly checks whether data is ready (a form of polling), followed by a blocking second stage to read the data.

3.3 IO Multiplexing

IO multiplexing allows a single thread to monitor multiple sockets using mechanisms such as select , poll , and epoll . The thread blocks only until at least one descriptor becomes ready, enabling concurrent handling of many connections.

3.4 AIO

AIO, introduced in Java 1.7 as part of NIO 2.0, is fully asynchronous: after issuing a system call, the user thread can continue working. The OS notifies the application when data is ready. Java provides two ways to handle the result:

Implement a CompletionHandler callback.

Use a Future object, checking isDone() and retrieving the result with get() .

3.5 Summary

BIO, NIO, and IO multiplexing all involve a blocking wait in the second stage and are therefore considered synchronous models, while AIO performs both stages without waiting and is classified as asynchronous.

Readers are invited to leave comments for further discussion.