How does the 'resume download' function work?

[do you know george?]

Hi all,

I have the DownThemAll! (download manager) extension installed into Firefox, and it allows you to Pause/Resume downloads for servers that allow it.

But does anyone know how the 'resume' function works?

Like, how does the server that you're downloading from KNOW that you have a partially completed file there; and how do they allow you to resume mid-download?

I've been trying to think up of some possible answers but they are probably ridiculous and improbable:

- Ridiculous Theory #1: The server itself stores a partially completed file there with your name on it for your convenience to resume later.

- Ridiculous Theory #2: Your Download Manager, reading the partially downloaded file on your hard disk, simply goes to the URL again, but somehow only resumes it from where you left off.

Hope someone can shed some light on this .

 

[markr]

When you do a http GET request for a file you can tell it which part of the file you want using a range header to the request. So your client knows how much of the file you currently have (just by look at the size of your current download) and it asks for the rest. I'm guessing this is how "threaded downloads" work as well, it sends 4 different GET requests each with a different range header.

 

[markr]

Very roughly a normal http GET to grab bigfile.zip from www.example.com might look like:
GET /bigfile.zip HTTP/1.1
Host: www.example.com

But instead you could do

GET /bigfile.zip HTTP/1.1
Host: www.example.com
Range: bytes=500-999
To only get that part of it. You can do this manually by telneting in to port 80 of a webserver.

 

[do you know george?]

Cheers mate thanks for that explanation, that was very informative.