Ahmed's Blog

Backend Communication Fundamentals: Day 2 Notes

 
  • Mar 30, 2025
  • 0 views

Backend Communication Fundamentals: Day 2

Section 1: Synchronous vs. Asynchronous Programming

Key Concepts

Synchronous Execution

  • Definition: A process is blocked while waiting for a response.
  • The caller sends a request and cannot proceed until it gets a response.
  • Comparable to a sine wave where client and server are “in sync”.
  • Example: A program requests the OS to read from disk → it gets removed from the CPU → resumes when data is ready.

Asynchronous Execution

  • Definition: The caller can continue execution while waiting for a response.
  • The request initiates an operation, and execution proceeds without blocking.
  • Two primary ways to check for a response:
    1. Polling methods (e.g., epoll in Linux)
    2. Callback mechanisms (e.g., io_uring in Linux, I/O completion ports in Windows)
  • Example: Sending an email while continuing other work, instead of waiting for an immediate reply.

Section 2: Real-Life Analogies

Comparing Synchronous and Asynchronous Workflows

Scenario Synchronous Asynchronous
Asking a question You wait for an answer before moving on. You send an email and continue working.
Chat apps Typing and waiting for a reply. Sending a message and getting notified later.

Section 3: Implementation in Node.js

Asynchronous Patterns in Node.js

Callbacks

readFile("large.dat", function onReadFinished(file) {
  // Handle file data
});
doWork(); // Execution continues immediately

Promises

readFile("large.dat").then(function (file) {
  // Handle file data
});

Async/Await

const file = await readFile("large.dat");
// Looks synchronous but is non-blocking

Section 4: Asynchronous Processing in Backend Systems

Queue-Based Processing

  • Backend servers often use message queues to handle asynchronous tasks.
  • A client request can immediately return a job ID while processing continues in the background.
  • The client can check status later or get notified when processing is complete.

Methods for Checking Completion

  1. Polling - Client repeatedly checks for updates.
  2. Push notifications - Server notifies the client when done.
  3. Long polling - Client makes a request that remains open until data is available.
  4. Pub/Sub systems - Clients subscribe to topics and get notified.

Section 5: Database-Specific Asynchronous Operations

Asynchronous Commits (PostgreSQL)

  • Synchronous Commit: Waits until data is written to disk (ensures durability).
  • Asynchronous Commit: Returns success before the data is committed (faster but riskier).

Asynchronous Replication

  • Synchronous Replication: Primary server waits for replicas to confirm before committing.
  • Asynchronous Replication: Primary commits instantly, and replicas update later (higher performance, lower consistency).

Section 6: OS-Level Considerations

File System Operations

  • OS often caches writes and flushes them asynchronously for performance.
  • Databases override this behavior using fsync to ensure durability.
  • I/O Event Handling Mechanisms:
    • epoll - Event-driven I/O notification.
    • io_uring - Asynchronous completion-based I/O handling.

Section 7: Key Vocabulary

  • Context Switching: Moving processes in/out of CPU execution.
  • Blocking: Execution pauses until an operation completes.
  • Event Loop: Node.js loop that handles callbacks, timers, etc.
  • Worker Threads: Additional threads used for blocking tasks.
  • Write-Ahead Log (WAL): Database journal of changes.
  • fsync: Command to bypass OS cache and write directly to disk.
  • Polling: Repeatedly checking if data is ready.
  • Callbacks: Functions executed when an operation completes.
  • Promises: Objects representing eventual completion of async operations.
  • Queues: Data structures for managing background workloads.
Search