Computational Storage Video Encoding for Host Bottleneck Relief
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video data management systems struggle with efficient storage and processing of video data across multiple storage devices, leading to bottlenecks and increased load on the host processor due to the need for decoding and retrieval of clips, which delays other commands.
Innovation Solution
Utilizing computational storage devices with accelerators to decode clips and optimize their retrieval and decoding, distributing clips across different computational storage units to reduce the load on the host processor and avoid data transfer between storage devices and the host processor.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If video clips are stored and managed using traditional file systems on multiple storage devices, then storage capacity is increased, but the host processor load increases and processing efficiency decreases
Solution Approach 1:
The patent introduces computational storage units as intermediary devices between the host processor and storage devices. These units include accelerators that perform video decoding operations, acting as mediators that offload processing tasks from the host processor while managing video clip storage and retrieval.
Solution Approach 2:
The patent extracts the decoding function from the host processor and places it in dedicated accelerators within computational storage units. This separation allows the host processor to focus on higher-level tasks while the accelerators handle the computationally intensive decoding operations.
2Ease of operation
If video clips are retrieved and decoded by the host processor, then data access is simplified, but the host processor becomes a bottleneck and other commands are delayed
Solution Approach 1:
Computational storage units serve as intermediaries that handle video clip retrieval and decoding operations. The accelerators within these units perform decoding locally, eliminating the need for the host processor to wait for decoding operations and preventing command delays.
Solution Approach 2:
The patent segments the video processing workflow into distinct functions: the file system manages clip locations, computational storage units handle retrieval and decoding, and accelerators perform the actual decoding. This segmentation allows parallel execution of multiple operations without bottlenecking the host processor.
3Quantity of substance
If multiple storage devices are used to store video clips, then storage capacity is increased, but predicting file locations and managing files becomes more complex
Solution Approach 1:
The computational storage units provide a universal interface for managing video clips across multiple storage devices. The accelerators within these units can handle decoding for clips from any storage device, creating a multi-functional system that manages diverse storage resources through a unified approach.
Data Source
Figure 1
Figure 2
Figure 3A
AI summary
A storage system (145) is disclosed. The storage system (145) may include a first storage device (120-1) and a second storage device (120-2). A receiver may receive an encoded stream. A splitter (610) may identify a first chunk (630-1) of data in the encoded stream and a second chunk (630-2) of data in the encoded stream. A distributor (615) may store the first chunk (630-1) of data on the first storage device (120-1) and the second chunk (630-2) of data on the second storage device (120-2).