Distributed Transcoding Using Idle Storage Server CPUs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network audio and video transcoding methods face issues such as large time consumption, low efficiency, high hardware costs, and poor scalability, particularly due to serial transcoding and the need for specialized hardware, which complicates resource management and scalability.
Innovation Solution
A transcoding method and system that utilizes idle computing resources on storage servers for parallel transcoding of media segments, reducing the need for specialized hardware and allowing for scalable cloud transcoding, where transcoding is performed directly on storage machines, minimizing data transmission and leveraging distributed file systems for efficient resource management.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If serial transcoding is used on dedicated video transcoding servers, then transcoding quality is maintained, but transcoding time is excessively long and efficiency is low
Solution Approach 1:
The patent divides the media file into multiple segments and distributes transcoding tasks for these segments across multiple storage servers simultaneously. Each storage server independently transcodes its assigned segment in parallel, transforming a single long serial transcoding process into multiple shorter parallel processes, thereby significantly reducing total transcoding time while maintaining quality
Solution Approach 2:
The patent enables storage servers to perform dual functions: both data storage and transcoding operations. By utilizing idle CPU resources on storage servers, the system eliminates the need for dedicated transcoding hardware, reducing costs while improving efficiency through parallel processing across multiple servers
2Productivity
If dedicated video transcoding servers are deployed, then transcoding quality is ensured, but hardware costs increase significantly
Solution Approach 1:
Storage servers are designed to perform both storage and transcoding functions simultaneously. By utilizing idle CPU computing power on storage servers during off-peak hours, the system eliminates the need for separate dedicated transcoding hardware, significantly reducing hardware investment while maintaining high transcoding efficiency through parallel processing
Solution Approach 2:
The system uses idle computing resources on storage servers themselves to perform transcoding operations. Rather than requiring external dedicated transcoding servers, the storage infrastructure serves its own transcoding needs by utilizing unused CPU capacity during low I/O demand periods
3Productivity
If multiple file transmissions are performed during transcoding, then data is transferred between servers, but transmission time and network overhead increase
Solution Approach 1:
The patent combines storage and transcoding operations into a single integrated process on the same server. By performing both functions on storage servers, the system eliminates the need for multiple separate file transmissions between dedicated storage and transcoding servers, reducing network overhead and transmission time while maintaining efficient parallel processing
4Productivity
If specialized hardware is used for segmentation and transcoding, then processing capability is improved, but device complexity and cost increase
Solution Approach 1:
The patent uses general-purpose storage servers with standard CPU resources to perform transcoding operations, eliminating the need for specialized transcoding hardware. The same storage infrastructure that holds media files also processes them, simplifying the overall system architecture while maintaining high processing capability through parallel operations across multiple servers
Data Source
AI summary
A transcoding method used in a computer network comprises: receiving, by a task manager in the computer network, a transcoding task, where the transcoding task has task information; generating task dispatch information according to the task information; and separately acquiring, according to the task dispatch information, source data of media segments corresponding to a media file, parallelly transcoding the source data into data in a target format by using multiple transcoders, and storing the data into a storage server in the computer network, where the transcoder includes a processor of the storage server. Distributed transcoding is completed by using an idle CPU of a storage server, so that not only costs of hardware are reduced and an existing resource is fully used, but also transmission and copying of data are greatly accelerated, thereby improving transcoding efficiency.