Using parallel stream allow running code in deparallel in a relatively simply way.

Configuration

Java Compilation:
Java Runtime:

JDK 11.0.12
JRE HotSpot 11.0.12

In Java, if there are several operations that the running sequence does not matter, we could create a thread for each operation to run them in parallel to improve performance

However, to create and use threads directly there are a lot of boilerplate code. Using parallel stream will achieve the goal of running code in parallel in a relatively simply way

First let's look into how long would it take to call 3 web urls sequentially:


public static void main(String[] args) {

    Date startTime = new Date();

    HttpClient httpClient = HttpClient.newBuilder().build();

    HttpRequest httpRequest = HttpRequest.newBuilder().uri(URI.create("http://www.nfl.com")).build();
    String response = "";

    try {
        HttpResponse<String> httpResponse = httpClient.send(httpRequest, HttpResponse.BodyHandlers.ofString());
        response = httpResponse.body();
    }
    catch (Exception e) {
        e.printStackTrace();
    }

    httpRequest = HttpRequest.newBuilder().uri(URI.create("http://www.nba.com")).build();

    try {
        HttpResponse<String> httpResponse = httpClient.send(httpRequest, HttpResponse.BodyHandlers.ofString());
        response = httpResponse.body();
    }
    catch (Exception e) {
        e.printStackTrace();
    }

    httpRequest = HttpRequest.newBuilder().uri(URI.create("http://www.mlb.com")).build();

    try {
        HttpResponse<String> httpResponse = httpClient.send(httpRequest, HttpResponse.BodyHandlers.ofString());
        response = httpResponse.body();
    }
    catch (Exception e) {
        e.printStackTrace();
    }

    Date endTime = new Date();

    System.out.println("Time taken: " + (endTime.getTime() - startTime.getTime()));
}

It will take:


Time taken: 904

We can make use of parallel stream to speed up the process

First create an ArrayList and add each operation arguments. Next call the parallelStream method to create a parallel Stream. Then define the same calling logic inside the map method. Finally call the collect method to get the responses as a list:


public static void main(String[] args) {

    Date startTime = new Date();

    List<String> list = new ArrayList<>();
    list.add("http://www.nfl.com");
    list.add("http://www.nba.com");
    list.add("http://www.mlb.com");

    List<String> listResponse = new ArrayList<>();

    listResponse = list.parallelStream().map(strURL -> {

        HttpClient httpClient = HttpClient.newBuilder().build();

        HttpRequest httpRequest = HttpRequest.newBuilder().uri(URI.create(strURL)).build();
        String response = "";

        try {
            HttpResponse<String> httpResponse = httpClient.send(httpRequest, HttpResponse.BodyHandlers.ofString());
            response = httpResponse.body();
        }
        catch (Exception e) {
            e.printStackTrace();
        }

        return response;

    }).collect(Collectors.toList());

    Date endTime = new Date();

    System.out.println("Time taken: " + (endTime.getTime() - startTime.getTime()));
}

This time it takes much less time:


Time taken: 431