The peer-to-peer file sharing systems have became very popular and widely used, in the recent years. However, these systems could not respond well to the increasing demand due to scalability and efficiency issues. Such problems become more obvious as the size of the file or the number of downloads in these system increases. In other words, current systems cannot perform efficiently under high loads.

To overcome these problems, we developed a novel algorithm that can be used to disseminate big files, like multimedia downloads, efficiently.

The algorithm schedules BitTorrent downloads to increase the speed of overall download times for its users. The important feature of the system is that the download time decreases as the number of concurrent number of download increases. Therefore, by managing the independent downloads the system improves the scalability whereas same situation creates an important performance loss in most of the current systems.

Moreover, the system makes the download of unpopular files, which exist in the system for a short period of time, available to a wider group of users.

In order to demonstrate the efficiency of this new method a large-scale flow-based simulator was implemented.