Maximizing users' utility for data transfer with priority control

Over the past few years, content delivery services using the Internet have been noted as a new business. In this type of data delivery, we consider services providing multiple classes having different data waiting times based on priority control. The utility to all users is thought to change with the number of classes under priority control and the traffic load ratio per class. Therefore, in this paper, we quantify the relationship between the waiting time and user utility in data transfers with priority control by employing a utility measurement function, and study the traffic load ratio of each class when provided with priority classes and maximizing the utility for all users. Then we analyze the conditions for maximizing utility for each data transfer method and the user utility measurement function, and demonstrate the optimum network efficiency through numerical examples. The results clearly show that a method providing multiple classes rather than a single class and using priority control increases the utility for all users. With priority control, the maximum utility is not reached at 100% network efficiency, but the maximum utility results when a few lines are free.