Adaptive Routing Algorithm for Joint Cloud Video Delivery

摘要

As the Internet keeps growing, online video has become a great part of the current Internet data traffic, which will take over 80% of Internet traffic according to Ciscos report. Also, new and more heavyweight applications keep developing to fulfill people's growing requirements, like 4k resolution and visual reality videos. However, one single service provider, like a Content Delivery Network (CDN), can not meet the performance requirements completely. To employ the potential of Joint Cloud, this paper designs and implements a new request routing algorithm that can make video delivery utilize multiple clouds and servers. On the premise of guaranteeing the quality of video playing, this algorithm minimizes the cost of service resources based on different infrastructures service quality, cost, and cover areas. Based on this algorithm, we implement a practical video delivery system using light-weight, flash-based terminals. And this system provides live video and video-on-demand delivery service for China Future Networks Industry Summit 2014 on June 4th. The actual user data was gathered and analyzed to verify the effectiveness of this algorithm. 90.2% of the total VOD requests were completed smoothly without pause, and the video traffic was optimized by the algorithm.