[AWS] Slow HTTP-upload performance

[Emmanuel Briot]

> I think there is should be user callback to receive uploaded data.
> Otherwise we would have to write uploaded data to temporary files, and
> user would have to move it somewhere else.
> Direct write to necessary place is at least 2 times faster.

I do not understand this. Moving a file, on almost all file systems, is 
almost free in terms of cpu and disk usage. And it is an atomic 
operation, which has the advantage that the file is only available in 
its final destination when it has been fully transfered.
I think the current implementation makes sense.

I would be interested to hear what the original poster meant by
"As I can see from the code, there is room for improvements". This seems 
to imply there are obvious speed ups, which would indeed be nice to put 
in place...

regards
Emmanuel