Node.js

Table of Contents

Sl.No	Questions
01.	Node.js Working
02.	Event loop working
03.	Flow chart of How does a node.js code works?
04.	Thread Testing
05.	Node's Asynchronous Feature
06.	Multitasking: Putting everything together
07.	Enhancing node performance
08.	The Callback Function
08.	Promises

Node.js Working

• V8 dependency: open source JS engine created by google, to execute JS code outside of the browser. V8 translates js stuffs into c++.
• Libuv Dependency: C++ open source project gives node access to the OS underlying file system, netorking also handles some kind of concurrency.
• Purpose of NodeJS: gives us interface to relate our JS side of our application to the actual c++ that’s running on our computer interpret and execute our JS code. It also provides a series of wrappers and a very unified inconsistent API for us to use inside our projects.
• Lib folder keeps all the Javascript implemntations.
• Src folder keeps c++ implementation of all the functions.
• Process.binding() connects JS and C++ functions.

Event loop working

• A Process is an instance of computer programm that is being executed.
• Within a single process we can have multiple threads.
• Thread has some instructions that has to be run by cpu (one by one).
• If we have more than one core inside our cpus we can process multiple threads at a time.
• Whenever you start a node code it automatically creates a thread and then executes the code inside that one thread.
• Event loop is like a control structure, decides what a thread should be doing at a given point of time.
• Node event loop is single threaded however some of Node framework std lib function that we run outside of event loop are not single thread.

Below is the pseudo code to explain the working of Event loop inside node.js

// node myFile.js

const pendingTimers =[];
const pendingOSTasks =[]; // all the things that OS async feature is handling
const pendingOperations = []; // All the things threadpool is hanling

// New timers, tasks, operations are recorded from myFile running
myFile.runContents();

function shouldContinue(){
  // check one: ANy pending setTimeout, setInterval, setIntermediate?
  // check two: Any pending os tasks? (like server listening to port)
  // check three: Any pending long running operation? (like fs module)

  return pendingOSTasks.length || pendingOperations.length || pendingTimers.length
}

// entire body executes in  one 'tick'
while(shouldContinue()){ // event loop
  // 1.) Node looks at pendingTimers and sees if any fucntions are ready to be called.
  // looks at setInterval or setTimeout functions.
  // 2.) Node looks at pendingOSTasks and pendingOperations and calls relevent callbacks.

  // 3.) pause execution. Continue when some no of events occur (like a new pendingOSTasks or pendingOperations are done or a timer is about to complete)

  // 4.) Look at pendingTimers. Call any setIntermediate.

  // 5.) Handle any 'close' event.
}
// exit back to terminal

Flow chart of How does a node.js code works?

Refer to further topics for better understanding

Thread testing

const crypto = require('crypto');
  
  const start = Date.now();
  crypto.pbkdf2('a', 'b', 100000, 512, 'sha512', ()=>{
    console.log('1:', Date.now()- start);
  });
  
  crypto.pbkdf2('a', 'b', 100000, 512, 'sha512', ()=>{
    console.log('2:', Date.now()- start);
  });

Output: 1: 1001, 2: 1012
  this output should have been 1:1001, 2:2000(approximately) if node was actually single threaded.

• This happens because libuv c++ side decides to run some computational functions ouside the event loop. They use ThreadPool(series of four threads to run comuptational tasks.)
• So four threads handle offloading work of expensive calculations outside the event loop.
• But if you use 5 pbkdf2 function ouptut will be like 1: 1750 2:1778 3:1781 4:1784 5:2661
• We can write custom JS codes that uses the thread pool.
• All 'fs' module functions use the threadpool.
• If you see above figure 2 functions are being assigned to a single core that's why it is taking more time to complete task 1,2,3,4 than in two functions case as cpu core has to do twice as work. When all these four function get completed 5th function gets added to thread pool. Below image explains the same thing.
• We can change threadpool size using process.env.UV_THREADPOOL_SIZE = 2;

Node's Asynchronous Feature

• It makes all the requests in same time.

  const https = require('https');
   
   const start = Date.now();
   
   function doRequest(){
   https.request('https://www.google.com', res =>{
     res.on('data', () =>{});
     res.on('end', ()=>{
       console.log(Date.now()-start);
     })
   })
   .end();
   }
   
   doRequest();
   doRequest();
   doRequest();
   doRequest();
   doRequest();
   doRequest();

    OUTPUT : 334 341 342 348 376 382 (NOT USING THREADS).
  

• Instead libUV delegates the request making to the underlying operating system, So it's actually our operating system that does the real HTTP request libUV is used to issue the request and then it just waits on the operating system to emit a signal that some response has come back to.

• Everything around netorking uses underlying OS, os's async feature

• Node.js is non-blocking I/O, non-blocking means it continues to do other tasks while it is renering ata from some task. One task doesn't block other.

• Asynchornous programming node.js keep doing other tasks while another task is taking time in fetching some data refer below:

  console.log('Starting');
  
  setTimeout(()=>{
    console.log('2 sec timer');
  }, 2000)
  
  setTimeout(()=> {
    console.log('0 sec timer');
  }, 0)
  
  console.log('Stopping');
  
  OUTPUT: 
  Starting
  Stopping
  0 sec timer
  2 sec timer
  `

• Call Stack keeps all the running functions starting from main function in a stack.

• Why does 0 seconds timer is printing after 'Stopping' log because functions like setTimeout use Node's API to proceed when such funtions get added to call stack they move to Node's API where they complete their time/task and move to Callback queue inside Event loop but we cannot add task from callback queue to Call stack until call stack is empty, And now our Call stack has main function in it, our code will proceed ahead and print 'Stopping' log, the log will be removed and after that the main function will be removed from stack, now stack is empty, So it will start taking functions from Callback Queue.

  

Multitasking: Putting everything together

If we put all the codes together as

	const https = require('https');
   const crypto = require('crypto');
   const fs = require('fs');
   const start = Date.now();
   
   function doRequest(){
   https.request('https://www.google.com', res =>{
     res.on('data', () =>{});
     res.on('end', ()=>{
       console.log(Date.now()-start);
     })
   })
   .end();
   }
   
   function doHash(){
   crypto.pbkdf2('a', 'b', 100000, 512, 'sha512', ()=>{
     console.log('Hash:', Date.now()- start);
   });
   }
   doRequest();
   
   fs.readFile('multitasks.js', 'utf8', ()=>{
     console.log('FS:', Date.now()-start);
   });
   
   doHash();
   doHash();
   doHash();
   doHash();

OUTPUT: 
340
HASH: 1583
FS: 1585
HASH:1596
HASH:1598
HASH:1609

• If we remove all the hash functions from the above code the file reading from hard drive (fs.readfile) shows output FS:28(very fast) but if we add hash functions then what's happening here?
• Everything inside FS module uses Thredpool. But still how does 5 thread(threadpool has 4 threads only) calls are showing same time, that's beacause when FS fuction goes to thread it make request to Hard drive for file access and then thread frees the FS function until the HD request gets completed and asign that thread to remained HASH function and when one of any hash function gets completed that spot gets asinged to FS function and as we know FS function complete quickly, So that's why one HASH function got completed before FS but if we increase our threadpool size then FS will completed quickly. And if we use 1 then all the hash functions will complete beforeb fs.

Enhancing node performance

• Two ways to improve node performanc. a. Use Node in 'Cluster' mode b. Use worker Threads.

• Blocking Event loop:

  const express = require('express');
  
  const app =express();
  
  function doWork(duration){
    const start= Date.now();
    while(Date.now()-start<duration){}
  }
  
  app.get('/', (req, res)=>{
    doWork(5000);
    res.send('Hellp, World!');
  });
  
  app.listen(3000);

• doWork function contains a while loop which runs for 5 seconds(does nothing), since event loop is single threaded, so our app will do nothing for 5 seconds(or will take 5 secons to load), It means one single tiny function is blocking our event loop. Since our event loop is blocked, So,if we try to run our app in two different tabs one after another, later will wait for first request to complete, See the time difference in below image.

• Clearly effect of very long running or computational intensive code inside a node project is not ideal at all.

Clustering

• Clustering is used to handle event blocking situation. Below diagrams explain the working of clustering.

  

  const cluster=require('cluster');
  
  // console.log(cluster.isMaster);
  
  // Is the file being executed in master mode? Checking if we have to go to cluster manger or if we are coming from cluster.fork() (Explained in above figure) and we have to go to Worker instance.
  if(cluster.isMaster){
  //Cause index.js to be executed again but in child mode
  cluster.fork();
  cluster.fork();
  cluster.fork();
  cluster.fork();
  }else{
  
  //Im a child I'm going to act like a server and do nopthing
  const express = require('express');
  
  const app =express();
  
  function doWork(duration){
  const start= Date.now();
  while(Date.now()-start<duration){}
  }
  
  app.get('/', (req, res)=>{
  doWork(5000);
  res.send('Hellp, World!');
  });
  
  app.get('/fast', (req, res)=>{
  res.send('Hellp, Flash World!');
  });
  app.listen(3000);
  }

• Above we used multiple cluster.fork() and created a new route '/fast' and if we try to run 'localhost:3000' and localhost:3000/fast simultaneously, we will see that localhost:3000/fast runs quickly while localhost:3000 takes 5 seconds(since it has a function call which runs for five seconds). So that's how clustering avoids event loop blocking. And if we use just one cluster.fork() then again both routes will take some time to process.

The Callback Function

• If we look at function below

const add =(x, y)=>{
 setTimeout(()=>{
   console.log('Two sec later')
   return (x+y)
 }, 2000)
}


const sum=add(1, 4)
console.log(sum);

OUTPUT: 
undefined
Two sec later

We see that above function is showing undefined output in place of sum of two numbers, that's because what has been explained in Call stack working section of Node's Asynchronous Feature but how can we get the get the sum here is using The Callback Function: A callback is a function called at the completion of a given task; this prevents any blocking, and allows other code to be run in the meantime.

const add =(x, y, callback)=>{
 setTimeout(()=>{
   callback(x+y)
 }, 2000)
}


add(1, 4, (sum) => {
   console.log(sum) // Should print: 5
})

OUTPUT
5

• Callback chaining example:

geocode(location, (error, {latitutde, longitude, location} ={})=>{
  if(error){
    return console.log(error);
  }
  forecast(latitutde, longitude, (error, forecastData) => {
    if(error){
      return console.log(error);

    }
    console.log(location)
    console.log(forecastData)
  })
})

Promises

It allows to associate handlers to an asynchronous action's eventual success value or failure reason. This lets asynchronous methods return values like synchronous methods: instead of the final value, the asynchronous method returns a promise for the value at some point in the future.

Promises in node.js promised to do some work and then had separate callbacks that would be executed for success and failure as well as handling timeouts. Another way to think of promises in node.js was that they were emitters that could emit only two events: success and error.The cool thing about promises is you can combine them into dependency chains (do Promise C only when Promise A and Promise B complete).

The core idea behind promises is that a promise represents the result of an asynchronous operation. A promise is in one of three different states:

pending - The initial state of a promise. fulfilled - The state of a promise representing a successful operation. rejected - The state of a promise representing a failed operation. Once a promise is fulfilled or rejected, it is immutable (i.e. it can never change again).

• Node.js promise expample

const doWorkPromise= new Promise((resolve, reject)=>{
  setTimeout(()=>{
    // resolve([1, 6, 7])
    reject('Things went wrong')
  },2000)
})

doWorkPromise.then((result)=>{
  console.log(result);
}).catch((e)=>{
  console.log('Error!', e);
})