I think that these download accelerators work by opening multiple connections to a site and perform byte-range fetching on blocks of the file, then stitch it back together.
You could easily do this with HttpClient and a few threads. There are zero copy writing to file examples on the website... I would start there and try to modify that example. Bill- On Mar 31, 2012 4:13 AM, "tommmmmm" <[email protected]> wrote: > I was searching for some java API for multi-threaded downloading and I came > across jdownman which heavily uses HttpClient. However that project is long > gone and google doesn't return any other sane results, so I decided to ask > here. How do you do multi-threaded/parallel/chunk downloading with > HttpClient? To be precise: something as for example IDM does or Flashget > does (or other so called download accelerators). > > I can easily download file with .copyUrlToFile function, however that is > single-stream download. I would appreciate help on this. Either as with > HttpClient itself or pointing me to some project/API that works ( I know > jDownloader supports multi-threading downloading, but the code there is > such a black magic to me and so complicated I can't figure heads and tails > of it, so importing the jDownloader.jar won't do me any good...) > > Thank you for your help. > > Tom > > > ps1: Please do not suggest (as at some `other` place) that most websites > limit me to 1 connection so it's useless to write such project and you > shouldn't want it. I fully understand which websites do allow more than 1 > connection and this project is to work with them delibarately. > ps2: Installing aria2c is not an option (as any other software) because I > do not have root access. I only have 'wget'. >
