As a software developer, you certainly have a high-level picture of how web apps work and what kinds of technologies are involved: the browser, HTTP, HTML, web server, request handlers, and so on.

In this article, we will take a deeper look at the sequence of events that take place when you visit a URL.

1. You enter a URL into the browser

It all starts here:

image

2. The browser looks up the IP address for the domain name

 image

The first step in the navigation is to figure out the IP address for the visited domain. The DNS lookup proceeds as follows:

  • Browser cache – The browser caches DNS records for some time. Interestingly, the OS does not tell the browser the time-to-live for each DNS record, and so the browser caches them for a fixed duration (varies between browsers, 2 – 30 minutes).
  • OS cache – If the browser cache does not contain the desired record, the browser makes a system call (gethostbyname in Windows). The OS has its own cache.
  • Router cache – The request continues on to your router, which typically has its own DNS cache.
  • ISP DNS cache – The next place checked is the cache ISP’s DNS server. With a cache, naturally.
  • Recursive search – Your ISP’s DNS server begins a recursive search, from the root nameserver, through the .com top-level nameserver, to Facebook’s nameserver. Normally, the DNS server will have names of the .com nameservers in cache, and so a hit to the root nameserver will not be necessary.

Here is a diagram of what a recursive DNS search looks like:

500px-An_example_of_theoretical_DNS_recursion_svg 

One worrying thing about DNS is that the entire domain like wikipedia.org or facebook.com seems to map to a single IP address. Fortunately, there are ways of mitigating the bottleneck:

  • Round-robin DNS is a solution where the DNS lookup returns multiple IP addresses, rather than just one. For example, facebook.com actually maps to four IP addresses.
  • Load-balancer is the piece of hardware that listens on a particular IP address and forwards the requests to other servers. Major sites will typically use expensive high-performance load balancers.
  • Geographic DNS improves scalability by mapping a domain name to different IP addresses, depending on the client’s geographic location. This is great for hosting static content so that different servers don’t have to update shared state.
  • Anycast is a routing technique where a single IP address maps to multiple physical servers. Unfortunately, anycast does not fit well with TCP and is rarely used in that scenario.

Most of the DNS servers themselves use anycast to achieve high availability and low latency of the DNS lookups.

3. The browser sends a HTTP request to the web server

image

You can be pretty sure that Facebook’s homepage will not be served from the browser cache because dynamic pages expire either very quickly or immediately (expiry date set to past).

So, the browser will send this request to the Facebook server:

GET http://facebook.com/ HTTP/1.1
Accept: application/x-ms-application, image/jpeg, application/xaml+xml, [...]
User-Agent: Mozilla/4.0 (compatible; MSIE 8.0; Windows NT 6.1; WOW64; [...]
Accept-Encoding: gzip, deflate
Connection: Keep-Alive
Host: facebook.com
Cookie: datr=1265876274-[...]; locale=en_US; lsd=WW[...]; c_user=2101[...]

The GET request names the URL to fetch: “http://facebook.com/”. The browser identifies itself (User-Agent header), and states what types of responses it will accept (Accept and Accept-Encoding headers). The Connection header asks the server to keep the TCP connection open for further requests.

The request also contains the cookies that the browser has for this domain. As you probably already know, cookies are key-value pairs that track the state of a web site in between different page requests. And so the cookies store the name of the logged-in user, a secret number that was assigned to the user by the server, some of user’s settings, etc. The cookies will be stored in a text file on the client, and sent to the server with every request.

There is a variety of tools that let you view the raw HTTP requests and corresponding responses. My favorite tool for viewing the raw HTTP traffic is fiddler, but there are many other tools (e.g., FireBug) These tools are a great help when optimizing a site.

In addition to GET requests, another type of requests that you may be familiar with is a POST request, typically used to submit forms. A GET request sends its parameters via the URL (e.g.: http://robozzle.com/puzzle.aspx?id=85). A POST request sends its parameters in the request body, just under the headers.

The trailing slash in the URL “http://facebook.com/” is important. In this case, the browser can safely add the slash. For URLs of the form http://example.com/folderOrFile, the browser cannot automatically add a slash, because it is not clear whether folderOrFile is a folder or a file. In such cases, the browser will visit the URL without the slash, and the server will respond with a redirect, resulting in an unnecessary roundtrip.

4. The facebook server responds with a permanent redirect

image

This is the response that the Facebook server sent back to the browser request:

HTTP/1.1 301 Moved Permanently
Cache-Control: private, no-store, no-cache, must-revalidate, post-check=0,
      pre-check=0
Expires: Sat, 01 Jan 2000 00:00:00 GMT
Location: http://www.facebook.com/
P3P: CP="DSP LAW"
Pragma: no-cache
Set-Cookie: made_write_conn=deleted; expires=Thu, 12-Feb-2009 05:09:50 GMT;
      path=/; domain=.facebook.com; httponly
Content-Type: text/html; charset=utf-8
X-Cnection: close
Date: Fri, 12 Feb 2010 05:09:51 GMT
Content-Length: 0

The server responded with a 301 Moved Permanently response to tell the browser to go to “http://www.facebook.com/” instead of “http://facebook.com/”.

There are interesting reasons why the server insists on the redirect instead of immediately responding with the web page that the user wants to see.

One reason has to do with search engine rankings. See, if there are two URLs for the same page, say http://www.igoro.com/ and http://igoro.com/, search engine may consider them to be two different sites, each with fewer incoming links and thus a lower ranking. Search engines understand permanent redirects (301), and will combine the incoming links from both sources into a single ranking.

Also, multiple URLs for the same content are not cache-friendly. When a piece of content has multiple names, it will potentially appear multiple times in caches.

5. The browser follows the redirect

image

The browser now knows that “http://www.facebook.com/” is the correct URL to go to, and so it sends out another GET request:

GET http://www.facebook.com/ HTTP/1.1
Accept: application/x-ms-application, image/jpeg, application/xaml+xml, [...]
Accept-Language: en-US
User-Agent: Mozilla/4.0 (compatible; MSIE 8.0; Windows NT 6.1; WOW64; [...]
Accept-Encoding: gzip, deflate
Connection: Keep-Alive
Cookie: lsd=XW[...]; c_user=21[...]; x-referer=[...]
Host: www.facebook.com

The meaning of the headers is the same as for the first request.

6. The server ‘handles’ the request

image

The server will receive the GET request, process it, and send back a response.

This may seem like a straightforward task, but in fact there is a lot of interesting stuff that happens here – even on a simple site like my blog, let alone on a massively scalable site like facebook.

  • Web server software
    The web server software (e.g., IIS or Apache) receives the HTTP request and decides which request handler should be executed to handle this request. A request handler is a program (in ASP.NET, PHP, Ruby, …) that reads the request and generates the HTML for the response.

    In the simplest case, the request handlers can be stored in a file hierarchy whose structure mirrors the URL structure, and so for example http://example.com/folder1/page1.aspx URL will map to file /httpdocs/folder1/page1.aspx. The web server software can also be configured so that URLs are manually mapped to request handlers, and so the public URL of page1.aspx could be http://example.com/folder1/page1.

  • Request handler
    The request handler reads the request, its parameters, and cookies. It will read and possibly update some data stored on the server. Then, the request handler will generate a HTML response.

One interesting difficulty that every dynamic website faces is how to store data. Smaller sites will often have a single SQL database to store their data, but sites that store a large amount of data and/or have many visitors have to find a way to split the database across multiple machines. Solutions include sharding (splitting up a table across multiple databases based on the primary key), replication, and usage of simplified databases with weakened consistency semantics.

One technique to keep data updates cheap is to defer some of the work to a batch job. For example, Facebook has to update the newsfeed in a timely fashion, but the data backing the “People you may know” feature may only need to be updated nightly (my guess, I don’t actually know how they implement this feature). Batch job updates result in staleness of some less important data, but can make data updates much faster and simpler.

7. The server sends back a HTML response

image

Here is the response that the server generated and sent back:

HTTP/1.1 200 OK
Cache-Control: private, no-store, no-cache, must-revalidate, post-check=0,
    pre-check=0
Expires: Sat, 01 Jan 2000 00:00:00 GMT
P3P: CP="DSP LAW"
Pragma: no-cache
Content-Encoding: gzip
Content-Type: text/html; charset=utf-8
X-Cnection: close
Transfer-Encoding: chunked
Date: Fri, 12 Feb 2010 09:05:55 GMT

2b3
��������T�n�@����[...]

The entire response is 36 kB, the bulk of them in the byte blob at the end that I trimmed.

The Content-Encoding header tells the browser that the response body is compressed using the gzip algorithm. After decompressing the blob, you’ll see the HTML you’d expect:

<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN"   
      "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" 
      lang="en" id="facebook" class=" no_js">
<head>
<meta http-equiv="Content-type" content="text/html; charset=utf-8" />
<meta http-equiv="Content-language" content="en" />
...

In addition to compression, headers specify whether and how to cache the page, any cookies to set (none in this response), privacy information, etc.

Notice the header that sets Content-Type to text/html. The header instructs the browser to render the response content as HTML, instead of say downloading it as a file. The browser will use the header to decide how to interpret the response, but will consider other factors as well, such as the extension of the URL.

8. The browser begins rendering the HTML

Even before the browser has received the entire HTML document, it begins rendering the website:

 image

9. The browser sends requests for objects embedded in HTML

image

As the browser renders the HTML, it will notice tags that require fetching of other URLs. The browser will send a GET request to retrieve each of these files.

Here are a few URLs that my visit to facebook.com retrieved:

  • Images
    http://static.ak.fbcdn.net/rsrc.php/z12E0/hash/8q2anwu7.gif
    http://static.ak.fbcdn.net/rsrc.php/zBS5C/hash/7hwy7at6.gif
  • CSS style sheets
    http://static.ak.fbcdn.net/rsrc.php/z448Z/hash/2plh8s4n.css
    http://static.ak.fbcdn.net/rsrc.php/zANE1/hash/cvtutcee.css
  • JavaScript files
    http://static.ak.fbcdn.net/rsrc.php/zEMOA/hash/c8yzb6ub.js
    http://static.ak.fbcdn.net/rsrc.php/z6R9L/hash/cq2lgbs8.js

Each of these URLs will go through process a similar to what the HTML page went through. So, the browser will look up the domain name in DNS, send a request to the URL, follow redirects, etc.

However, static files – unlike dynamic pages – allow the browser to cache them. Some of the files may be served up from cache, without contacting the server at all. The browser knows how long to cache a particular file because the response that returned the file contained an Expires header. Additionally, each response may also contain an ETag header that works like a version number – if the browser sees an ETag for a version of the file it already has, it can stop the transfer immediately.

Can you guess what “fbcdn.net” in the URLs stands for? A safe bet is that it means “Facebook content delivery network”. Facebook uses a content delivery network (CDN) to distribute static content – images, style sheets, and JavaScript files. So, the files will be copied to many machines across the globe.

Static content often represents the bulk of the bandwidth of a site, and can be easily replicated across a CDN. Often, sites will use a third-party CDN provider, instead of operating a CND themselves. For example, Facebook’s static files are hosted by Akamai, the largest CDN provider.

As a demonstration, when you try to ping static.ak.fbcdn.net, you will get a response from an akamai.net server. Also, interestingly, if you ping the URL a couple of times, may get responses from different servers, which demonstrates the load-balancing that happens behind the scenes.

10. The browser sends further asynchronous (AJAX) requests

image

In the spirit of Web 2.0, the client continues to communicate with the server even after the page is rendered.

For example, Facebook chat will continue to update the list of your logged in friends as they come and go. To update the list of your logged-in friends, the JavaScript executing in your browser has to send an asynchronous request to the server. The asynchronous request is a programmatically constructed GET or POST request that goes to a special URL. In the Facebook example, the client sends a POST request to http://www.facebook.com/ajax/chat/buddy_list.php to fetch the list of your friends who are online.

This pattern is sometimes referred to as “AJAX”, which stands for “Asynchronous JavaScript And XML”, even though there is no particular reason why the server has to format the response as XML. For example, Facebook returns snippets of JavaScript code in response to asynchronous requests.

Among other things, the fiddler tool lets you view the asynchronous requests sent by your browser. In fact, not only you can observe the requests passively, but you can also modify and resend them. The fact that it is this easy to “spoof” AJAX requests causes a lot of grief to developers of online games with scoreboards. (Obviously, please don’t cheat that way.)

Facebook chat provides an example of an interesting problem with AJAX: pushing data from server to client. Since HTTP is a request-response protocol, the chat server cannot push new messages to the client. Instead, the client has to poll the server every few seconds to see if any new messages arrived.

Long polling is an interesting technique to decrease the load on the server in these types of scenarios. If the server does not have any new messages when polled, it simply does not send a response back. And, if a message for this client is received within the timeout period, the server will find the outstanding request and return the message with the response.

Conclusion

Hopefully this gives you a better idea of how the different web pieces work together.

148 Comments to “What really happens when you navigate to a URL”

  1. Shouvik says:

    As you told – “One worrying thing about DNS is that the entire domain like wikipedia.org or facebook.com seems to map to a single IP addres”

    I could not understand how two domains can have same IP address. Would not this create ambiguity while resolving DNS?

  2. Anonymous says:

    Nice article. very informative. I would have loved if you had written a bit about the roles of TCP and IP in the protocol stack.

  3. van der Zeen says:

    2 Shouvik:

    The author actually meant IP addressess are different, but DNS lookup returns the only IP for each of these domains. As you can see from the article, the returned IP can be different in some cases depending on load balancing, geo IP, etc.

    Nevertheless, there’s tons of situations when a number of domains resolved to the single IP, and that’s no problem for DNS, as forward DNS lookup is resolving a domain name to an IP address. Ambiguity will arise in the case of reverse DNS lookup, when IP address is resolved to a domain name. However it’s outside of this article topic, as a web browser performs only forward lookups.

    Web server can handle multiple domains with no problem basing on request sent by a browser (see Host field of the request).

  4. Sanjaya says:

    Great post. Just what i need to create the power point presentation for my students. thanx. btw when different domains point to the same IP then it is the web servers responsibility to present the correct domain files to the browser.

  5. aldwin says:

    Thanks for the informative and well written article Igor. I love the way you outline the steps clearly and connected the dots, even having those side notes in boxes. Definitely useful for beginners and for experienced developers like me who need a refresher once in a while :)

  6. varsha says:

    Great article. Very helpful! Thanks a lot.

  7. Anonymous says:

    good post…
    really helpfull

  8. flexbee says:

    Great article mate..! Gud job..!

  9. Betsy says:

    Great post – I do have a question though, if I have a Facebook page that has a URL with “3rdtimesthecharm”, but people might instead write “thirdtimesthecharm” in the URL…can I have multiple URLs for people to type in and have them both direct to my page?

    Thanks so much!

  10. Sanjeev says:

    Very Informative and crisp.

  11. web4betsy says:

    Betsy:

    Research redirects for redirecting various URI to a specified URI.

    The modrewite módule for Apache will provide a richer method to handle various URI/URLs. One can rewrite ‘thirdtimesthecharm’ to ‘3rdtimesthecharm’ using a regular expression, ‘*timesthecharm’ (research regular expressions; my syntax “*…” is simplified).

    Webserver rediects are simpler to setup but do not handle regular expression without a module.

  12. Karthik Krishnan says:

    Very good article… Good job

  13. Yannick says:

    Thank you very much!

  14. oxygen says:

    GET http://facebook.com/ HTTP/1.1
    should have been
    GET / HTTP/1.1

  15. oxygen says:

    False information:
    “The GET request names the URL to fetch: “http://facebook.com/”.”

    Should have read:
    The GET request names the relative URL to fetch.

  16. Anonymous says:

    Good one………. :-)

  17. Dhara says:

    Nice Article very interesting keep posting similar ones

  18. Deepak says:

    thank you
    It helps to understand about server response time.

  19. Manoj says:

    I was searching for a good article, yours seems to be the best so far. A visual display to users along with explanation is really good. Great job

  20. Ajay says:

    indeed a good article.really helpful

  21. Very Very Interesting and very Helpful. I learnt what is happening is the background of Internet when I type in a website name and actually hit enter on the keyboard.

  22. RogerLiu says:

    Thanks Ostrovsky,I would like to translate your article to Chinese if you allow me to do that.

  23. getster says:

    thank you, we explained

  24. vishal k says:

    Nice artical sir, this will help me give presentation in my class.
    Thanks………..

  25. Anonymous says:

    you make me understand permanent redirect, thank you

  26. Asawari says:

    Too Good Article. Explained very Nicely!!

  27. Praveen says:

    I got clear idea, Thanks.

  28. poor credit says:

    What i don’t understood is in fact how you’re no longer really much
    more well-appreciated than you may be right now. You are so intelligent.
    You know thus considerably in terms of this topic, produced me in my view imagine it from a
    lot of varied angles. Its like women and men are not interested unless
    it is something to do with Woman gaga! Your personal stuffs nice.
    Always handle it up!

  29. Anonymous says:

    only today i understood how this network concept works after 5 years of work experience

  30. Califrisco Navi says:

    Thank you so much for one of the most detailed reviews of this process that I’ve found. Thank you for all the work that went into this. Superb job Igor! :)

  31. G1 says:

    This was a helping fundaa… It helped a lot to see the hidden procedure that happens behind the screen…. thnx a lot to the author..

  32. Paul B says:

    For a request to “http://www.facebook.com/xyz.htm”:

    1) The “http:” bit tells the browser to use the http protocol, but it does NOT appear anywhere in the GET request

    2) the first line of the request will read:

    “GET /xyz.htm HTTP/1.1″
    which is the absolute reference to that page (note that if you had clicked on a “relative” link in a web page, your browser will have converted it to an absolute reference)

    3) the request will also have a line (often the second line, but it can be anywhere) with the fqdn in it:

    “Hostname: http://www.facebook.com
    (with HTTP version 1.1 this is a required header!)

    This “Hostname” header is needed as many servers now host many fqdn’s on a single IP address… so this tells it WHICH server the traffic is for.

    (FQDN = Fully Qualified Domain Name)

  33. shilong says:

    Good article,thank you!

  34. shilong says:

    Hi, if I can translate it into Chinese?

  35. Ronak says:

    Can I fetch the request before it is sent to ISP/Router via some programming in C++/C#?

  36. Cenk says:

    Thank you so much, very helpful tutorial.

  37. Naren says:

    Nice explanation… But still Can someone tell the details considering all the Lower layers of the networking model and also considering Digital certificates ??

    Complete explanation of what happens when we enter a URL in the address bar narrowed down to physical layer and ALL THE ELEMENTS THAT COME IN WAY.

    Regards,
    Naren

  38. Taher says:

    Excellent article but I am wondering>>>>
    Do these processes unique for different browsers???????

    Thanks

  39. Magnificent goods from you, man. I’ve understand your stuff previous to and you’re just extremely great.

    I actually like what you’ve acquired here, certainly like what you’re
    stating and the way in which you say it.
    You make it entertaining and you still care for to keep it sensible.

    I cant wait to read far more from you. This is actually a tremendous site.

  40. gamezebo.com says:

    We absolutely love your blog and find many of your post’s to
    be exactly what I’m looking for. Do you offer guest writers to write content for you?
    I wouldn’t mind creating a post or elaborating on some of the subjects you write
    related to here. Again, awesome site!

  41. kay28point5 says:

    Very nicely written article thanks.
    Was asked this question and was writing a detailed process overview myself, but wanted to check it against other input to see where my mistakes or omissions were. [Although I am looking for even more detail] this was very helpful. Thanks!

  42. Michael Wan says:

    Excellent article, just some comments.
    1. consider protocol, htttp/https
    2.content access authentication, need login? account have the permission?

  43. Anmol says:

    Hi

    Nice article giving high level view.

    Thanks

  44. Anmol says:

    Hi

    Can you give a similar article on how e-commerce website works. Like selecting items to cart and then order them.
    What are the different steps and the communication between client and server.

    Thanks

  45. S Bandi says:

    Great description of what happens behind the browsing.

  46. Santosh says:

    Hi,

    What will happen if i type request in the browser (url) and close the tab?
    Below are my queries:
    1. Will the server executes the request even if requester is no more exists?
    2. Does server keeps track of requester? (Whether it is alive or not)?

  47. Alok kumar says:

    It is very useful article.
    I really love to read such article in future too…

Leave a Reply

You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>