Question

I’m developing a simple crawler for web pages. I’ve searched an found a lot of solutions for implementing multi-threaded crawlers. What is is the best way to create a thread-safe queue to contain unique URLs?

EDIT:
Is there a better solution in .Net 4.5?

Answer 1

Use the Task Parallel Library and use the default scheduler which uses ThreadPool.

---

OK, this is a minimal implementation which queues 30 URLs at a time:

    public static void WebCrawl(Func<string> getNextUrlToCrawl, // returns a URL or null if no more URLs 
        Action<string> crawlUrl, // action to crawl the URL 
        int pauseInMilli // if all threads engaged, waits for n milliseconds
        )
    {
        const int maxQueueLength = 50;
        string currentUrl = null;
        int queueLength = 0;

        while ((currentUrl = getNextUrlToCrawl()) != null)
        {
            string temp = currentUrl;
            if (queueLength < maxQueueLength)
            {
                Task.Factory.StartNew(() =>
                    {
                        Interlocked.Increment(ref queueLength);
                        crawlUrl(temp);
                    }
                    ).ContinueWith((t) => 
                    {
                        if(t.IsFaulted)
                            Console.WriteLine(t.Exception.ToString());
                        else
                            Console.WriteLine("Successfully done!");
                        Interlocked.Decrement(ref queueLength);
                    }
                    );
            }
            else
            {
                Thread.Sleep(pauseInMilli);
            }
        }
    }

Dummy usage:

    static void Main(string[] args)
    {
        Random r = new Random();
        int i = 0;
        WebCrawl(() => (i = r.Next()) % 100 == 0 ? null : ("Some URL: " + i.ToString()),
            (url) => Console.WriteLine(url),
            500);

        Console.Read();

    }