Vector Processing Unit Transcoding via Queue Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Telecommunication systems face inefficiencies in transcoding large amounts of audio data due to the need to convert between different media formats, which consumes significant processing resources, especially when handling thousands of simultaneous communication sessions.
Innovation Solution
A method and system that utilize a network node with queues for transcoding packets between different communication formats, where packets are processed by a vector processing unit, such as a graphics processing unit, to efficiently convert data between various formats by placing them into vectors and applying series of functions for transcoding.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional processing units are used to transcode audio data between different communication formats, then transcoding functionality is provided, but processing resources are significantly consumed and processing efficiency is low
Solution Approach 1:
The patent segments the transcoding processing into multiple queues (first queue, second queue, third queue) that can be processed independently and in parallel. Each queue handles specific packet formats or communication sessions, allowing the processing unit to work on multiple segments simultaneously, thereby improving overall productivity while distributing resource consumption across multiple processing streams.
Solution Approach 2:
The patent introduces a queue-based dimensional structure to organize and process packets. By placing packets from different communication sessions and formats into separate queues, the system adds a processing dimension that enables parallel handling of multiple transcoding tasks, significantly improving processing efficiency without proportionally increasing resource consumption.
2Productivity
If large quantities of audio data are transcoded simultaneously for thousands of communication sessions, then communication capacity is maintained, but processing time increases and call quality may deteriorate
Solution Approach 1:
The patent performs preliminary actions by organizing packets into queues before actual transcoding processing. Packets are pre-categorized into different queues based on their format and communication session, so that when processing begins, the system can immediately start parallel processing without additional organization overhead. This preliminary organization reduces overall processing time while maintaining high throughput.
Solution Approach 2:
The patent ensures continuous processing by maintaining multiple queues that can be processed in parallel and by implementing a system where the processing unit continuously draws from available queues. This continuity eliminates idle time between processing tasks, ensuring that productive action is maintained throughout the transcoding process, thereby reducing total processing time while handling large data quantities.
Data Source
Figure 1
Figure 2
Figure 3A~3B
AI summary
A method includes, with a computing system, receiving a plurality of packets for transcoding, the plurality of packets comprising a first set of packets having data in a first format and a second set of packets having data in a second format, with the computing system, placing the first set of packets in a first queue, the first queue comprising packets from more than one communication session, with the computing system, placing the second set of packets in a second queue, the second queue having packets from more than one communication session, with the computing system, processing packets of the first queue with a graphics processing unit, and with the computing system, after processing the packets of the first queue with the graphics processing unit, processing packets of the second queue with the graphics processing unit.