I think I figured it out. Here is v4.
If the number of requests is less than 1 GB, the algorithm stays the sameas before. If we need to process more, we will do it incrementally withslices of 1 GB.
Best regards,Maxim Orlov.