On file delay minimization for content uploading to media cloud via collaborative wireless network

This paper investigates the problem of uploading user-generated content files (e.g., video captured on mobile devices) to a media cloud via a cooperative wireless network. The multi-path opportunity inherent to the cooperative network provides a new dimension by optimally allocating all the packets into different paths to improve the user experience. Specifically, we aim to minimize the end-to-end file delay, resulting in a better experience for end users. We first show that the file delay is penalized by a path-starvation effect, which results from poor packet allocation among different paths. Secondly, our in-depth analysis of the canonical two-path case indicates that it could hurt in some cases to use more paths. Especially, when the delay variance of assigning one packet to the slower path is on the same order of that of assigning all the packets to the faster path, using one path is better than using both paths. Finally, based on this insight, we propose an iterative algorithm to assign packets into a set of chosen paths to alleviate the path starvation problem, with an objective to reducing the end-to-end file delay. Numerical simulations demonstrate its near-optimal performance. The insights from our theoretical analysis could provide guidelines for platform and application development.