Video On Demand - High Performance Cost-Effective Model
Abstract
Video streaming service is a substantial portion in the present day’s world of the internet and is exploding. Studies have predicted that videos will account to 81% of the internet traffic by 2021 [32]. The most challenging research for the past decade is performed on how to build a cost effective distributed VoD architecture while maximizing the quality of experience for the end users. After doing a study on how to improve the performance of VoD systems, we have identified transcoding, storage, caching, load balancing, and CDN integrations as the major contributing areas for providing users a better quality of experience. Video streams usually require conversion, also known as transcoding depending on the device the client uses to stream live or on-demand videos. However, this is an expensive option to transcode all the videos because of the high computation, storage, and retrieval involved in the management of these videos. Due to the constant increase of network traffic in the network backbone, caching plays a significant role in helping the overall system by storing the multimedia content nearer to the end user. To balance the load across the network while transmitting the video it is important to maintain constant metrics like bit rate, movements of frames. By combining VoD with CDN, high quality of service can be provided to the users with less bandwidth cost and maximum throughput. Various methods from the recent studies are identified and explained in this paper for each of these sections, and based on that we have proposed a hybrid model that combines all of these approaches to provide a high performance and cost-effective design for VoD servers.
Keywords
VoD (Video-on-demand), Transcoding, Storage, Caching, Load balancing CDN, cost-effective, performance