System for high performance on-demand video transcoding
Inventors
Bayoumi, Magdy A. • Li, Xiangbo • Salehi, Mohsen Amini
Assignees
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Publication Number
US-11166057-B2
Publication Date
2021-11-02
Expiration Date
Abstract
Video streams, either in form of on-demand streaming or live streaming, usually have to be transcoded based on the characteristics of clients' devices. Transcoding is a computationally expensive and time-consuming operation; therefore, streaming service providers currently store numerous transcoded versions of the same video to serve different types of client devices. Due to the expense of maintaining and upgrading storage and computing infrastructures, many streaming service providers recently are becoming reliant on cloud services. However, the challenge in utilizing cloud services for video transcoding is how to deploy cloud resources in a cost-efficient manner without any major impact on the quality of video streams. To address this challenge, in this paper, the Cloud-based Video Streaming Service (CVSS) architecture is disclosed to transcode video streams in an on-demand manner. The architecture provides a platform for streaming service providers to utilize cloud resources in a cost-efficient manner and with respect to the Quality of Service (QoS) demands of video streams. In particular, the architecture includes a QoS-aware scheduling method to efficiently map video streams to cloud resources, and a cost-aware dynamic (i.e., elastic) resource provisioning policy that adapts the resource acquisition with respect to the video streaming QoS demands. Simulation results based on realistic cloud traces and with various workload conditions, demonstrate that the CVSS architecture can satisfy video streaming QoS demands and reduces the incurred cost of stream providers up to 70%.
Core Innovation
The invention provides a cloud-based video streaming service (CVSS) architecture for on-demand cloud transcoding. Video streams are split into independently transcoding segments, each including at least one Group of Pictures (GOP). The GOP segments are treated as deadline-bearing tasks based on a deadline derived from first-frame presentation time, with close-GOP assumed.
A QoS-aware transcoding task scheduler maps interleaved GOP tasks to homogeneous transcoding virtual machines with a local queue. The scheduler prioritizes GOPs in a startup queue while ensuring batch-queue deadline feasibility and does not assume prior knowledge of an arrival pattern of Groups of Pictures to the system. Local scheduling is performed within the transcoding virtual machines using FCFS.
An elasticity manager performs dynamic (event- and periodic) VM provisioning using threshold-based deadline-miss-rate control and predicts a deadline miss rate that occurs at a subsequent provisioning event to decide scale-up and scale-down. A lightweight remedial resource provisioning policy is applied based on startup-queue size, and a video merger reconstructs output streams. A caching policy avoids unnecessary transcoding for long-tail access patterns.
Claims Coverage
The independent claim covers a CVSS system with seven major components: video splitter, transcoding task scheduler, transcoding virtual machines with local queue, elasticity manager, video merger, video repository, and caching policy. The coverage centers on splitting video into independently transcoding GOP segments, QoS-aware scheduling without prior knowledge of GOP arrival patterns, and elasticity that predicts a subsequent provisioning event deadline miss rate.
On-demand cloud-based transcoding system with splitter and scheduler
A system comprising a video splitter, a transcoding task scheduler, at least one transcoding virtual machine, an elasticity manager, a video merger, a video repository, and a caching policy, wherein the video splitter splits video streams into segments with at least one Group of Pictures that can be individually transcoded, and wherein the segments are transcoded with one said Group of Pictures.
Local queue on transcoding virtual machine
The at least one transcoding virtual machine comprises a local queue.
Scheduling without arrival-pattern knowledge
The transcoding task scheduler does not comprise prior knowledge of an arrival pattern of Groups of Pictures to the system.
Elasticity manager with dynamic resource provisioning policies
The elasticity manager comprises at least one dynamic resource provisioning policies.
Predicting subsequent deadline miss rate for provisioning
The dynamic resource provisioning policies comprise functionality to predict a deadline miss rate that will occur at a subsequent provisioning event.
Overall, the independent claim requires GOP-based segmentation for independent transcoding, a scheduler that does not rely on prior GOP arrival patterns, transcoding VMs with a local queue, and elasticity that uses dynamic resource provisioning policies to predict a subsequent provisioning-event deadline miss rate.
Stated Advantages
Provides a startup delay of less than 1s [as reported in the provided summary].
Provides an average deadline miss rate of less than 10% [as reported in the provided summary].
Provides up to 70% cost reduction [as reported in the provided summary].
Documented Applications
Used in a cloud-based video streaming service (CVSS) architecture for on-demand cloud transcoding with QoS-aware scheduling and elasticity, including caching for long-tail access patterns [as described in the provided summary].
Interested in licensing this patent?