I have spent quite a lot of time searching an answer for this question but haven't found an satisfactory answer.

My perception is that everything that you see in your browser is downloaded on your PC. The proof for this is: Let a webpage load in your browser and when its loading completes then select working offline or close your internet connection and save that page, which results in the saving of your webpage on your hard disk with the key to watch this page any number of time you want.

So, it shows that your page's been downloaded when you accessed it and the speed of accessing your webpage will surely then depend on your downloading speed, then where does this term "Surfing Speed" comes.

I may be wrong in my perception but I would love to know the correct answer to my query. Thanks. And yes I admit that my proof does not validate when the webpage is developed in languages like asp.

  • Another way to be sure about time your browser takes to render the page is to use Opera, and come back to previous page. The page will be loaded from cache only, without requesting connection (for most pages at least). – Gnoupi Jun 25 '10 at 12:13
  • @Gnoupi also, most browsers support the [CTRL]+[F5] shortcut to bypass the cache and reload the page 100% from the server – STW Jun 25 '10 at 16:53
  • related (dupe?): superuser.com/questions/89814/… – yhw42 Apr 16 '11 at 12:59

Download speed is how fast the connection is from your house to your ISP. It is tested by downloading a file hosted right on your ISPs network to your computer (all speedtests not hosted locally by your ISP are testing latency). It is the physical speed of your "last mile," between your house and your ISP's gateway to its backbone. Also known as throughput.

Surfing speed is a marketing term, and is meant to describe how fast the connection is, subjectively, from your house to items on the Internet. It is always variable, and is out of any one entity's control. It includes latency (as explained above, basically 'how long it takes information from servers around the world to cross the Internet and render on your machine'). Also known as goodput.

  • 1
    This is one of the better explanations I have seen that makes clear the difference between your connection to the internet (you to your ISP) and your connection to other servers (anything that goes through your ISP to other networks - google, yahoo). Most people do not make this distinction. The other issue is how one actually defines speed, and that is usually classified as latency (response time of servers over a connection). How many bits per second (throughput) is more of the size of the pipeline, not the speed of the connection. – MaQleod Mar 13 '11 at 21:19
  • I'll expand on this with an anology. Your destination to a server is similar to a route between your home and a geographical location. The latency of each hop is like the speed limit on each section of road, that is the fastest each hop will let a packet travel. Throughput is like the amount of lanes on the road. The fewer lanes and the more traffic there is, the higher the latency gets. When you have more lanes, traffic can flow more freely and can reach the maximum speed allowed. When you download too much, you clog your pipe and your speed lowers. – MaQleod Mar 13 '11 at 21:32
  • To sum it all up, surfing speed is essentially the latency of each hop to a destination server added together, plus the processing time your computer needs to render the information it receives. – MaQleod Mar 13 '11 at 21:33
  • Yup. Across the 'net and up the stack! – goblinbox Mar 15 '11 at 1:43

Downloading usually involves only one file, but surfing involves many (html, javascript, images etc.).
Normally it takes time for a download to reach its peak speed.

Therefore, downloading several small files is slower than one large file.
But not only because of the download speed, but also because of the handshake that needs to be done between the browser and the server in order to start a new download.

This can be optimized by the browser's reusing one connection to the server in order to download multiple files, but not all website administrators allow keeping connections alive, so as not too have too many concurrent connections open.

In conclusion:

Download speed is normally done at the maximum connection speed (at least after the first few seconds).

Surfing speed depends on :

  • The browser's efficiency in connecting only once to each web server linked-to on the webpage
  • The browser's efficiency in serving up the page even if not all the pieces have arrived yet
  • The website server's administrator allowing the keep-alive of connections
  • The foresight of the webpage writer in minimizing the number of separate items and their sizes

In conclusion I would say that while download speed is well-defined, surfing speed can vary greatly depending on the above parameters.

  • HTTP 1.1 can send multiple requests over the same connection, so there's only one handshake. – vtest Jun 25 '10 at 16:50
  • as mentioned in my answer, different browsers also support different numbers of concurrent connections--so older browsers may feel much slower because they allow fewer requests to run in parallel – STW Jun 25 '10 at 16:55
  • @vtest: That's exactly what I said. – harrymc Jun 25 '10 at 19:59

Tricky, depends what does the surfing term mean exactly.

If it's so called, "user experience", then let's take easy example (I hope) - modern websites are using on javascript. Scripts are downloaded from server with some speed depending on your badwith, but they are executed locally by your browser. Those scripts can do some additional computing: database queries/generating data/html content, which takes additional time.

Web browsers differ in javascript executing time. It depends on your system resources too.

So with the slow machine with old browser will get worse surfing speed even if it's provided with the same bandwith as the better machine.

That's how I suppose term 'surfing speed' should be understood. But who is using it? If it's your ISP talking, then consider it as a smooth excuse for them no providing you a quality connection (like: "our badwith rox, it's your machines that sux") :)

  • A thing to keep in mind also is that websites usually require data connections to several servers: the actual data you want to reach, maybe a server for images, then advertisements, a traffic meter, a Facebook social plugin, etc. Loading a webpage is then slower because of this repeated calls to several servers. – Gnoupi Jun 25 '10 at 12:14

Bad Latency can mean a connection which is good for downloading large files is bad for surfing.

Long time ago, before broadband was available here, we tried a internet via satellite service.

Download speed was fantastic (for it's time), so if you started to download a large file (eg a service pack or cd image), then it took a few seconds to get started and then it flew.

However, when you were browsing the internet, the same delay occurred for every section that you downloaded. A webpage contained five images means you are downloading six items and the delay occured six times. So the delays mounted up.

Hence the broadband connection was 20 (maybe more) times faster than dialup when downloading files, but for browsing the internet it was a lot slower.


Also, when you visit a webpage, all the content for that particular page can come from multiple servers across the world, like advertising, video's and such, so it depends on so many factors of how fast any particular webpage loads at any given time, regardless of your connection speed.


[copying my answer across from where you asked the same question on Stack Overflow]

Not necessarily the same thing. Your actual speed of downloading data will be constrained by the bandwidth of your network connection, so suppose you can get X kB / second then that's how many actual bytes of data you can slurp.

Now think about the things that affect how that actually appears to you:

how fast does your broswer render a downloaded page?

  • a webpage with a dodgy connection at the other end might not serve the page at the full rate that you could retrieve it at
  • the raw data rate includes actual TCP?IP packets, a webpage that has to go through 7 redirects to actually be displayed to you (with a DNS look up and a new TCP connection and HTTP request each time) will appear much more slowly
  • your ISP might offer compression, whereby you visit a webpage, the ISP compresses it at their end and then you run special decompression software at your end. So, fewer actual bytes of data to be transferred, but the compression / decompression time might outweigh that
  • size of your ISP's proxy cache - and is the page you requested already in it or not

All of these things taken into account will give you your "surfing speed". It's really a slightly meaningless number given that it varies wildly depending on exactly what you are looking at.


Surfing is more dependent on response time. Downloading has more to do with bandwidth. Suppose I have an internet connection that takes 150 milliseconds for a www.google.com to respond... If I have 50 of those internet connections, it'll still take 150 milliseconds, but I'll have 50 time as much bandwidth, so 50 times the downloading speed. Surfing still depends a lot on download speed, especially when getting pictures or flash or movies.

Also a lot of ISPs cache certain websites to make them go faster. This just means that the ISP which is pretty close to you holds a copy of a website hosted in a foreign country, you're request doesn't always have to go to that foreign country to get that website if your ISP is holding a local copy of it. This speeds up surfing speed as well, though the same principle could apply to download speeds if your ISP chooses to cache any of those too.


ISP favours real time communication traffic than non time sensitive traffic Surfing is real time and download non time sensitive


Your ISP may consider them to be very different things. It's not uncommon for ISP's to advertise a "burst-rate" which means that files download very quickly initially, but slow-down over the long run. The implication of this is that small files (such as HTML, JS, CSS, etc) can receive much higher speeds than large files (ISOs, zip, hi-res photos, etc).

Another implication is that your "bandwidth speed tests" may be skewed--if the test sends only small files then it may report higher bandwidth than you'll see transferring larger files--and vice-versa.

Another factor is "which client". Back in the dial-up days download accelerators were popular for retrieving large files over HTTP/FTP; older webbrowsers significantly limit the number of concurrent requests allowed in retrieving resources (CSS, images, etc) compared to modern browsers, and so on. So the software you're using can also impact the performance you observe.

Regarding caching; you are correct that your browser tries to store and reuse files it has already retrieved--this cuts down significantly on bandwidth and server load. Using an HTTP debugger like FireBug or Fiddler2 you'll notice many pages that you've already visited will respond with an HTTP StatusCode of 304 -- "Not Modified". What has happened is your browser sent the request, and in that request said "I need SomeResource, I already have version X of that" if the server sees that version X is the most recent then it responds with a status of 304--basically telling the browser "use what you already have" rather than resending the resource.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.