You can access parts of the ISO file past your quota size by creating a sparse file using the truncate command, then appending to it with "wget -c". For example, if you want to get the part of the file past 3,000,000,000 bytes, create a sparse file of this size with the same name as the ISO:

truncate -s 3000000000 Fedora-16-i386-DVD.iso

Since the file is sparse, it uses almost no actual disk space and doesn't count against the quota. Then you can append to it with wget -c:

wget -c http://dl.fedoraproject.org/pub/fedora/linux/releases/16/Fedora/i386/iso/Fedora-16-i386-DVD.iso

But even to do that I need to have at lease 3.5GB space on my fedorapeople.org space. Isn't it?

You would only need the amount of space taken up by the amount that wget downloads (since the sparse part doesn't count against the quota). Also, even if you increase your quota, if you download the entire DVD and then split it into smaller pieces, the pieces will take up about the same amount of space as the original file (although you could delete the original file immediately afterwards). So you would actually need a quota of about twice the size of the ISO, almost 8 GB.

From the users mailing list I see you're working on this (based on http://linuxandfriends.com/2008/11/01/curl-split-a-file-and-download-simultaneously-from-multiple-locations/ ). That's a better solution in the long run, since it's easy to automate and works for arbitrarily large files. You could also use it to manually download RPMs larger than 50 MB that you need for updating (for example, if libreoffice-core is updated, you couldn't use yum update, you'd have to download it manually in the same way as the ISO).

Actually, using yum-presto (which is installed by default in F16), libreoffice-core will probably be updated using a deltarpm which is much smaller than 50 MB. But if you want to install a new package, it will be a full download, and it may be larger than 50 MB. Better to learn how to use the curl script to download arbitrary files.

So, was a solution found to this? Or is this space still needed?

well, I couldn't check with it yet. But it seems that the method at http://linuxandfriends.com/2008/11/01/curl-split-a-file-and-download-simultaneously-from-multiple-locations/ will work. So no need of extra space.